No. 80 - June 1995 Recent Developments R. GIACCONI, Oirector General of ESO

Many significant events have occurred environment for astronomical purposes notae clarifying the procedures, by the in the past few weeks and months wh ich and on programmes of development of establishment of a point of liaison for ESO will affect the life of ESO foryears to come. Chilean astronomy and related technolo­ at high level in the Chilean Foreign The most important one has been the gies, will play an important role in our Ministry and by the liberation of all signing of the "Interpretative, Supplemen­ future cooperation. materials wh ich had been upheld at tary and Amending Agreement" to the In a joint press release the Govern­ customs. The issue of the claims of 1963 Convenlion between the Govern­ me nt of the Republic of Chile and the ownership of Paranal by the La Torre me nt of Chile and ESO, by Mr. Roberto European Southern Observatory ex­ family is being resolved by the Govern­ Cifuentes, Plenipotentiary Ambassador pressed their desire to continue to work ment of Chile in out of courts direct representing the Government of Chile towards the resolution of common discussions wh ich will hopefully lead to a and the Director General of the European problems in a spirit of mutual respect and settlement in a short time. The Govern­ Southern Observatory on April 18, 1995. full collaboration. The signing of the me nt of Chile has given ESO both public This Agreement, wh ich will have the Agreement occurred after months of and private assurances that Chile effect of widening and strengthening the dialogue between the parties and considers this problem as an internal cooperation between ESO and the constitutes an important step towards the Chilean issue wh ich should not affect Chilean scientific community, will be solution of some of the pending points on ESO-Chile relations. In any case, work on submitted for ratification by the National the current agenda for discussions Paranal is proceeding at full speed. Congress of the Republic of Chile and by between Chile and ESO. Among the This generally positive evolution of the the ESO Council. Upon taking effect, this remaining points the most important situation in Chile has been the result of Agreement will ensure guaranteed ac­ appear to be weil on their way to solution many efforts. The determined effort and cess by Chilean astronomers to all ESO thanks to the determined efforts of the resolve of the Chilean Government to facilities up to 10% of the observing time. Chilean Government. deal with the issues in a decisive manner It is expected that for VLTNLTI half of this The issue of ESO immunity of cannot be overemphasized. The actions time will be devoted to collaborative jurisdiction on Paranal wh ich was violated by the Ambassadors and Foreign efforts between Chilean and European on March 30, 1995bythevisitofa member Ministries of the member states of ESO in astronomers. In the field of labour of the Judiciary with the support of the c1arifying to the Chilean Government the relations, the Agreement will result in the police, has been dealt with an official concerns of the European states for the incorporation of so me fundamental expression of regrets by the Chilean difficult situation in wh ich various events principles of Chilean labour legislation in Government and assurances by the had placed the VLTNLTI project were the internal ESO rules and regulations for Government that all necessary steps essential. Finally, the continued and the local statt in Chile. Chilean scientists within the Chilean constitution would be patient efforts of many, many diplomats, named by a Chilean Committee will taken to prevent the recurrence of such scientists and administrators both Chile­ participate in all ESO scientific and incidents. an and European have been essential to technical committees. Joint Chile-ESO The issue of imports and accreditation progress. We at ESO are extremely committees on the preservation of the has been dealt with by the issuance of two grateful to all of them. The ESO Council, wh ich met in expect the date of first light to be aHected These past few months have been also Extraordinary Session on April 19, 1995, by 3 to 6 months. Financial losses have extremely significant in the development expressed its satisfaction for the many been experienced due to work stoppag­ of an even closer involvement of the positive steps which had been achieved, es, delays, increased costs for storage of Science Divisions both in Garching and its intent to ratify the Agreement and components, rescheduling of planned La Silla in all technical aspects of ESO instructed the Executive to continue the activities, etc. Resolution of these activities. Staff scientists and Fellows are direct negotiations with the Chilean financial issues will require considerable expected and do contribute to operations, Government to hopefully resolve all attention by the ESO Executive and maintenance and upgrading of existing remaining issues prior to the regular Council over the next year. facilities, to the development of new Council meeting of June 7 and 8. In the technical areas the meetings of instrumentation, to the development of Important and fundamental as the issue of the Users Committee and of the Scientific new software for proposal processing, Chile is to permit us to continue carrying Technical Committee have resulted in scheduling and implementation, to the out astronomical programmes, it was only important decisions regarding the future development of physical models of the one of significant events in the last few of La Silla. STC has approved the instruments, to trade-oH studies between months. construction of SOFI (a near infrared scientific requirements and engineering The first comprehensive VLT manage­ imaging spectrometer for the NTT) and difficulties, to development of detectors, ment report based on a Work Breakdown the beginning of the assessment phase of in short to all essential activities of ESO as Structure was presented to the ESO the 3.6-metre upgrade plan. The La Silla an observatory. While this involvement is Scientific Technical Committee on May 4 2000 group is completing its work of being strengthened by proceeding with and 5 and to the Finance Committee at its planning lorthe future of La Silla in the VLT the hiring to the budgeted staff level, the meeting of May 9 and 10. This report, era. The increased attention by ESO to cooperation between ESO scientists and wh ich will be issued every six months at the development and operation 01 optical engineers in Chile and Europe is the request of Council, was very warmly detectors with competitive quantum increasing. Slowly, the concept that ESO received by both committees. Apart from eHiciencies, speed of read-out and noise is a single observatory, with a single staff its intrinsic interest for those who are has al ready resulted in notable improve­ whether in Chile or Europe, is emerging. involved in the monitoring of the technical ments. A plan for the continuation 01 this Detector development projects are the and managerial aspects of the VLT eHort lor the VLT instrumentation was responsibility of people at La Silla as weil programme, it represents the culmination enthusiastically endorsed by both UC and as Garching, the NTT and the 3.6-m of a very intensive effort by many groups STC. relurbishment eHorts are being carried at ESO to restructure the accounting, The VLT Science Operation plan was out by mixed groups, software develop­ reporting and management information presented to the STC and received a very ment is being carried out jointly and so systems to cope with the challenge of the positive approval and recommendation to forth. We consider this approach essen­ VLT project. proceed even further in implementing the tial to the successful operation 01 fa­ In the February ESO-wide annual end to end approach to implementation 01 cilities such as the VLTIVLTI in the review the same management principles the science programme that it represents. future. were applied to the reporting of every In general, the increased attention to the I would like to conclude these brief ESO activity both in Europe and Chile. planning for the utilisation 01 the VLT notes by expressing my increasing The efforts of the VLT Division, the newly telescopes and instruments has resulted confidence that ESO and the European created Instrumentation Division, Admin­ in a first cut study for the necessary astronomical community will prove equal istration and Project OHice, and 01 the software and hardware tools that will need to the challenges of the next century. I Chile Administration were essential in to be developed. Increased attention to base this confidence on the evidence I making this possible. These new tools will and formal representation of the data f10w see of increased cooperation between be even more important to permit clarity of necessary to take us Irom proposal entry ESO and its Scientific and Technical communication between ESO manage­ to scheduling, data reception, calibration, Committee and the European astronom­ ment and its oversight committees in the reduction and archiving has shown the cial community. On the solidarity which diHicult linancial times that one can see considerable amount of work still in front was expressed in ditlicult times by all ahead. of uso Arecent ESO Workshop on member states. On the remarkable While the overall technical progress on Calibration and Data Management performance by the European hardware VLT is extremely encouraging, the Techniques was extremely successful in contractors. Finallyon the splendid diHiculties in Chile in the recent past have permilting us to measure our progress performance by the entire ESO staH resulted in both time and financiallosses with respect to past and current large everywhere and in every function. which are currently being evaluated. We projects in astronomy.

TELESCOPES AND INSTRUMENTATION News from the VLT Programme M. TARENGHI, ESO-Garching

This has been an incredibly challeng­ Programme and ESO as a whole. The 01 difficulty was the civil engineering work ing period for the VLT Programme. The most serious problems were related to being performed at the VLT site by the simultaneous difficulties in Chile and two diHiculties with the import of material and Joint Venture Skanska-Belfi wh ich was major external contracts have taxed the equipment to Chile and the accreditation aggravated by the Chile situation. Finally, management resources of both the VLT of contractor personnel. The second area the loss by Matra Marconi of access to a

2 Figure 1: Mirror No. 1 in the final stages of polishing. Figure 2: Mirror No. 2 in the grinding stage. crilical silicon carbide technology for the the first mirror is nearing completion at area with many technical risks, and two secondary mirror assembly of the VLT REOSC with the on-time completion 01 contractors had entered into competition. required ESO to negotiate a contract with the remaining mirrors projected. After a design competition between Zeiss Dornier, lhe second qualified bidder, The contract for the main structure is and GIAT, the GIAT design was found to resulting in both cost and schedule in­ also proceeding weil. The AES consorti­ fully meet the Programme requirements creases. um is scheduled to complete the unit and be both superior in performance and These problems have to date resulted one telescope subsystem by September lower in cost. The prototype produced by in a projected schedule delay, with 1995. Calculations indicate that the com­ GIAT indicates that the final system will reference to the June 1993 baseline of 5 pleted unit will deliver the best rejection of perform as required. months to first light. external disturbances ever obtained on a The secondary mirror M2 has been In spite of these problems, excellent lelescope of this size. The conlractor is mentioned as one of the costJschedule work has been performed in most 01 the scheduled to complete erection of the unit problem areas of the project. It should be major areas 01 the Programme. For telescope in Milan at the end 01 December pointed out, however, that the original example, the heart of the VLT, its 8.2­ 1995. The large mechanical structures for impact caused by the MATRA problem metre primary mirrors, are progressing Main Structure Units 2, 3 and 4 have (six months on first light) has now been weil (Figs. 1 and 2). The first two blanks already been started and are proceeding reduced to only three months' delay. The have been delivered by Schott and according to plan. contract with Dornier-DASA is proceed­ exceed expectations. The third and fourth The complex M1Cell-M3 tower, wh ich ing on schedule, although with a higher blanks are due in July and October 1995 supports the 8.2-metre primary mirror, price than ESO's original baseline project and are proceeding on schedule. It using 150 active supports which con­ price projections because of the use of should be remembered that this was stantly maintain the correct shape of the beryllium mirrors for the first unit. originally considered one 01 the greatest mirror, is entering the manufacturing The four telescope enclosures are also risk areas 01 the Programme. Polishing 01 phase. This was considered the second progressing weil. The contractor has sent

Figure 3: Paranal on 2 April 1995.

3 Figure 4: The completed seelion of the delay line tunnel. Figure 5: Inside the building of the first Unit Telescope.

8 shipments with over 300 tons of material new ESO CCO prototype system, ACE, Another positive development has to the site to begin erection as soon as the has been tested on La Silla and is been the selection of the VLT Programme importation and accreditation problems functioning weil. Scientist wh ich still has to be confirmed are solved. Work in Europe is proceeding A number of management changes by Council. The VLT Programme Scien­ as planned. have also been implemented in the tist will provide the nucleus for a small At the Paranal site, the work on the Programme. The VLT work was organ­ group of scientists to assist the VLT foundation for the first telescope is almost ised into work packages beginning in Programme in scientific issues. Also the complete (Figs. 3, 4 and 5). The total site February 1994 with budget planning by new Project Scientist for the VLTI, F. concrete work is 25% complete. Although work packages implemented with the Paresce, has begun work, adding senior significant problems with schedule have 1995 budget. In January 1995, the scientific oversight to this important plagued this contract, we believe that the accounting systems at ESO were modi­ Programme area. critical work for telescope No. 1 can be fied to include cost collection by work In February 1995, the third annual completed in May 1995 to allow the next package in addition to the traditional cost ESO Wide Review was held. In this contractor, SEBIS, to begin erection ofthe collection by nature of expense. Review, the work, schedule and cost for enclosure. It should be noted that the Another critical area was the system the year 1995 were reviewed. A key contractor has also been affected by the engineering. The vacancy of the head of element of this Review was the introduc­ import problems. the System Engineering Group has now tion of clear quarterly progress mile­ One of the critical areas of the been filled and the new head of system stones wh ich form the basis for control by performance of the VLT and its scientific engineering brings in-depth telescope the upper management. instruments is the CCO detectors. A new system experience to ESO and has now head of this group has been recruited. recruited personnel for the remaining This group has been strengthened with open system engineering positions in­ additional manpower to ensure that this cluding the important area of configura­ M. Tarenghi important area is properly covered. The tion control. e-mail: [email protected]

Scientific Priorities for La Silla in the VLT Era J. ANOERSEN, Chairman of the Working Group

In a previous article (The Messenger, years, and the facilities that will be The Questionnaire Survey No. 78, p. 3). the rationale behind the needed to do so. creation of the ESO Working Group Afteranalysing the replies, the WG has Nearly 150 replies were received by (WG) on Scientific Priorities for La prepared a first draft plan to serve as a mid-February, a quite respectable turn­ Silla in the VLT Era was outlined. At skeleton forthe following discussions. We out. While any such survey will inevitably the same time, the WG solicited the are grateful for the keen interest of the be both incomplete and biased for a views of the community on the main community and would like to present number of reasons, and cannot be an c1asses of science to be carried out belowa brief status report on ourworkand exact measure of the community's plans from La Silla over the next five to ten the plans for its completion. and wishes, this material is an invaluable

4 guide for our work. Also, several olf preparatory work is required for VLT ° ADONIS should stay on the 3.6-m thoughtful and pertinent comments were projects, the corresponding instrumenta­ until CONICA + adaptive optics enter made on a variety of individual topics. tion on La Silla must be available in time. operation on the VLT. Some main trends are al ready clear ° The VLT will completely outclass ° MEFOS and OPTOPUS lag behind from the answers: some current La Silla facilities. Yet, high­ contemporary efficiency by large factors °Strong demand for wide-field imaging priority projects must be done on La Silla and should both be retired. Competitive (> 1°). visible and IR. in the interim. successors cannot be completed early ° Strong emphasis on survey-type °New initiatives mustbe focused on the enough, and time exchange agreements work, both stand-alone and in preparation largertelescopes, which are both the main should be explored instead. for VLT projects. VLT partners and the most labour­ ° A higher spectral-resolution option ° Much demand for moderate- and intensive to run. For the smaller tele­ and efficient fiber link to the 3.6-m are high-resolution spectroscopy in the scopes « 1 m), the 1993 recommenda­ essential for the long-term competitive­ visible, moderate resolution in the IR. tions remain in force unless otherwise ness of the CES. Strong interest in a wide-field, multi-object stated. spectrograph (MOS) with > 400-500 °At all times, facilities must be planned fibers. to achieve maximum operational simpli­ Towards a Final Plan ° Demand for long-term monitoring of fication; this implies single-configuration variable sources, with requests to keep a telescopes and block scheduling of The draft was discussed at length at photometric telescope on La Silla (some instruments as far as possible. the UC and STC meetings in May. There for polarimetry as weil). Accurate ° The true financial impact of the was a gratifying measure of support in standard stars for the VLT must be proposed measures is not primarily in the both committees for the overall strategy established. direct costs, but in paving the way for a of the draft as weil as most individual ° The role of La Silla in hands-on more cost-effective organisation of La proposals. The natural wish of users to training of young astronomers is seen as Silla as a whole. maintain La Silla instruments in top shape very valuable, but second in priority to until their VLT successors are in stable excellent science. Draft Recommendations 01 the operation was stressed by both. The ° Users are emphatic that La Silla must WG educational role of La Silla, defined as remain internationally competitive; small A first rough timetable and list of student training, met with some skepti­ and medium-class telescopes continue to actions was distributed to the Users cism. The important issue is, however, have valid and valuable roles to play. Committee and STC in late April. It will no the experience of those scientists who, doubt undergo many revisions before a 20 years from now, will teach a new General Policy Considerations final version is reached, but some of its generation ofstudents. Hence, additional current main elements are the following: suggestions and comments will be In trying to chart the course of La Silla °A careful tradeoff study of mirror size, solicited in the next version of the report. into the future, we are guided by some of image quality, and ease of operation is After further discussion in the com­ the landmarks previously set. Two of needed to define the future home of wide­ munity, the WG plans to meet the UC, these are the report by the WG on field optical imaging on La Silla. Results STC, and OPC together for a final review Scientitic Priorities tor the VLT Obser­ so far indicate that the 3.6-m is unlikely to and refinement of the plan, and its vatory (1995) and that on Scientitic become a competitive facility. financial implications will have conse· Priorities tor La Si/la Operations (see The °Wide-field imaging and medium/low­ quences in the 1996 budget. Final pre­ MessengerNo. 74, p. 29,1993). We must resolution spectroscopy in the near IR will sentation to the DG and Council follows now carry the 1993 plans forward in mesh remain vital. The proposed NTT instru­ after the November STC meeting. with the VLT project, guiding La Silla to a ment SOFI will coverthese needs in a very In order to facilitate community steady-state situation after the year 2000. cost-effective way, and the WG recom­ access, later drafts of the report mayaiso Some of the basic premises for the mends that it be built. become available on the WWw. preparation of such plans are: ° Until VLT+VISIR take over, TIMMI ° The timetable of changes is driven should be upgraded with a larger array, primarily by the schedule of the VLT and even temporarily, especially for ISO Johannes Andersen its instrumentation. follow-up work. e-ma;/: [email protected])

Is the Seeing Situation at the 3.6-rn Telescope Irreversible?

M. FAUCHERRE, ESO-La Silla

1. Introduction monitor. This situation could be im­ The dome is so large (30 m diameter) that proved, and we are convinced now that residual sources of heat produce internal Image quality (10), or the sharpness in one year's time it would be possible to thermal gradients wh ich cannot be of the point spread function (PSF) at obtain 0.8" FWHM long exposures with eliminated by using the wind or any forced a 3.6-m instrument focus, should not fall EFOSC1 routinely. This would however dome ventilation from outside, just below one arcsec n FWHM when exter­ require a large effort during that period. because the dome cannot be opened nal conditions are excellent, and is worse Poor 10 at the 3.6-m is not due to the sufficiently - unlike the NTT dome. than 1.15" FWHM most of the time (see site itself or to the quality of the optics Besides, obtaining thermal equilibrium Fig. 1). Images are hardly ever as sharp (z 0.45" FWHM), but rather to "mirror with the outside all the time by means of as at the NTT, the 2.2-m or the seeing seeing" and to the presence of the dome. cooling and ventilation is not realistic,

5 Monthlyaveraged eeing (wavelength & zenith corrected) in the telescope environment, which would take into account changing 2.2 1"T"'-rt-r--r-r-~...... ,rrr-t-r"'--.-r-.-r""'-rr-]:::::!:!:!:!:!:±::r::!:!:I:::d:J!:!:!::!:!jr external conditions, so that the 10 is always optimal. Cooling systems either ---seeing 3.60m already in place or presently being 2 ... . ~ ~ ~ . - -. -- seeing 2.20m constructed (M1 surface cooling) will ü ...... Seeing Monitor include line TO adjustment capabilities lor ~ 1.8 critical items such as M1. ~ ~ "0 1.6 2. A Historical Perspective ~ ::J ~ 1.4 Twice in the past, a dome ventilation Q) E system was built and tested. No 10 improvement was noticed. Finally, in Ol 1.2 c 1990 it was decided to build an Air Q) Q) Conditioning (AirCo) and a Floor Cooling Cf) (FloorCo) system, following recommen­ dations by R. Le Poole, an engineer who 0.8 studiedthe problem in somedetail in 1990 (ref. 1). This system was operational in 0.6 October 1993 just belore a Come-On­ I'-

Vl'rlic;lllcIllIH'ralllrl' l:rndienl fccor

7 t-~"""~"'-~~"""T'"~"""~'-TH=-:j2!-:m-'-(T~r....,.~~\-TT"~~j'-T-.!-T""" E• J o "o ~ ::: ~:~:~:~:~~:I::: ~I:·f: :~:~::::1:::::: :; _ß ~ H=+riable!T w.~ . T'\~) : • (' .. , ( " .. ( ,.( : ; . ~ 4 o : H=~.5m(T; ;. T ,) , o "''''T~'~'~ .~~." I 3 .: ...... lOt ß I I• 2 E E .2'· ·er o iii J:; J:: 3 0 , .g .g· •> ·1 •> ·1

~ ·2 ~ ·2 , l • ·3 +-...... 1--'...... 1--'...... -1-...... + ...... +"""'-+ ·3 +-O'-t-...... ;...... -t-...... ;...... -t-...... ;...... -i-...... ~ j 14h 16h 18h 20h 22h 24h 2h 4h 6h 8h lOh 12h 14h j 14h 16h 18h 20h 22h 24h 2h 4h 6h 8h lOh 12h 14h TImeUT TImeUT Figure 2: Temperalures recorded at various heights in the dome ("half tube": on Serrurier truss in between M1 and top-ring, "floor": from fixed sensor inside false ftoor) in Oclober 1994. The seeing was better than 0.7" during both nights. White all T" were wilhin 0 1° the last night, leading to 5 1" FWHM PSFs, they were totally non uniform on October 19 (FWHM:i? 1.6").

6 March 23-24 : DIMM4 v. DIMM2 March 24-25 : DIMM4 v. DIMM2 March 16-17 : CCD v. DIMM2 1.04 ••.•• :.•••.••••••••..•••• ;...... ••.•••••.•: ..••..

1.7 ;::::I::::t::j:::::::::~:::::::::j::::r:::;:::':::::; 0.96 ____ ._._ ••••• _ •••••• 4. _ •••• : ••••••• : ••• , •••••• . ,, i·····~····~~··~·a·i·····~····~····~····~···· ~ ····i····~ : : cl' : : . 8 l.S : : :0 : : : : r : : : : :.: filll!II!ii] 00 : :::::;:::::J ::::':1:'····r:::··:r::'::. ;':': .::: .. :::: .E !····i····j····t·~····:····(·! ····!····!·..·(·t 1.3 ~ ~ o t 0 [ • l ::E r:::::;::~~:::::~:~::::::;.:::r::::;::::i:::T:::; e. 13 0.72 ·····:·······!······r·--·· ·······;······'i·······;.······ ::E 1.1 J: 6 ~ 1.1 t::!::::r {:.•::::r:::r:r:::r::!::::r::::

O.Q :.: ·r::::;.:::r:::r:::r::::\::::r:::f:·::[:·::I::::r::1 0.7 :::J······l···rr·····I······~······ 0.4 ",".4~0.e...6...... 0. S~6-0""'.6~4~0-'-.7~2~O.L.8~0.L.6~6~O.L.6....-U..04 0.7 0,9 1.1 1.3 1.5 1.7 0.7 0.9 1.1 1.3 1.5 1.7 0 4 Q 1 FWHM (DIMM2) in arelee FWHM (DIMM2) In are..e FWHM (OIMM2) in .rcl.c

March 20-21 : SHARP v. DIMM2 March 24-25 : SHARP v. DIMM2 Mlrch 19-20 : SHARP v. DIMM2 (." t.mperature. wlthln 0.9 deg.) 1. 8 r--.,.....("SI~R~IU,...S,_3"rm_•.,g~••_b,....o..;rP"rtl_on..;.)_~ : : ~ : ····;·····:······r·_·_·t·····1······f·····1·····r····· ... _···~······~·_·_·~·····j······~··-·-~···-·i······l..... ,-,­ _._~ ~ ~ _-~ ~ 1. 6 j.. \ : +.. ~:::::;:::: .... ~ ~ ~ ; ~ ?~ ] ~ ~ . 1.6 :::::::::::::::::j::::T:::L:::C::·.::.. '.' •••;.\t1 ••• 111 ••• I: ~ ,. ~ f 1. 4 ~ ....••~ •....•..... ~ ••••..:- ..•.•..•••i······:-···· . L···l···: + ,. 1.4 ····r···O[·····1·····T·····f·····~ , :: l l . ····1······[·····1···· ~ ····j·····i·..···(····t·····;·····r-····t·····(· .. ·f··· ····i······~o····j·····~······: ... _-! ..... ~ ..... +._._-! .... ···+·0··+···+····+·····: ~ ~ ~ + . 1.2 : :: : : j j ! i ···f~···!··~··t····- ~ ···~·····i·····i······!·····:···· ::::L::I::::··::·:r-::~::[:::··;:::::r:~:I::.:.::::: '2 .. :::L~r:::L:T:::I::::L:T::::[:::::L: ·····t·····i······~···öi·····~······r·····i···· , I 1 ····1·····; '.'li:il!I:! : :0 : : 1 : 1 : : ;·····~·····1·····~·····t .... ,·····~······:·····1·····r····T·····l·····r·····~··· .... ,·····f··~·ja~·~·····T···T···T····T···T··· ...... \.....!.....+...+... 0.8 0.6 O. 8 "'::-...... i...,....L--l~L...... L....L...... l...... :l 0.8 1.2 , .4 1.6 1.8 0.8 \.2 1.4 1.6 1.6 0.8 1.2 1.4 1.6 1.8

FWHM (DIMM2) In are••e FWHM (DIMM2) in arc••e FWHM (DIMM2) In .reue Figure 3: Seeing measurements made on several occasions in March 1995, with MI cooled to ambient T" (T" :rIO) and floor cooled to T., -2". Comparison with (he seeing monitor DIMM2 gives an estimate o( remaining M1 and dome seeing. underneath the false floor and the primary electronic closets inside), M1 P still with ambient air. Convective disturbances mirror have thermal time constants of remained warmer than ambient by;::: 1.5°. caused by a mirror being warmer than several days. That is why it was decided in July 1994 to ambient cause serious image degrada­ • Cool dome floor and volurne. Le cool M1 (see section 5). Now, how does tion: a factor of 0.35" / °C of P imbalance Poole suggests that we inject cold air the AirCo system work? Cold air is for a horizontal mirror (an average upwards into the dome so as to produce injected both into the floor and at the top between different authors). Mirror seeing a "bubble of cold air", wh ich should help of the dome during the day; only the image broadening is produced by a to reduce air turbulence in the open slit. FloorCo is maintained during the night. turbulent boundary layer. Mirror seeing • Ventilate dome using outside air. 20 The system which was built for the was identified as the major contributor to dome volumes per hour are needed. It dome is right from the "physics overall seeing in several conventional was shown later that the cost - and standpoint", even though the cold air (pre-1985) facilities (CFHT, AAT, UH 88). moreover maintenance expenses - for produced is not uniformly distributed, as It has been suggested by several authors such a system would be prohibitive, given can be seen in Figure 2. As a result of that a flow of air across the surface of a the size and the form of the dome. thermal stratification, long time constant mirror may substantially improve the • Coolthe oilin the drives andeliminate items cannot reach thermal equilibrium seeing. Thedifficulty inthe caseofthe 3.6­ small heat sources. This last item refers with their environment, because convec­ m mirror is that it will be impossible to especially to the Cassegrain cage. Still tion is inhibited. So, more ventilation is maintain a laminar flow over a distance of today, new instruments like ADONIS or required in the dome. However, the more than a metre in the open air. TIMMI are mounted in the cage with large system did not solve what was found to be Dome seeing occurs when the air in the electronic racks including power supplies the major issue: mirror seeing. For this, telescope beam up to the telescope slit is and other heat sources. two systems were conceived. They are described in section 5. Also insulation out of temperature equilibrium with 2.2. The "AirCo" System would help a lot to reduce cooling expen­ outside air. Once again, this effect is ses: Highly conductive materials should worse when the dome is warmer, All recommendations by Le Poole be used to avoid thermal radiation. It is producing convection in the beam were implemented, except two: forced suggested in particular that we wrap the through the slit. It has been shown that dome ventilation, for reasons given in the telescope top end in aluminium foil. most wavefront disturbances occur in the introduction, and M1 cooling. Indeed it slit area. This was very c1ear on pupil was assumed that with the combination of images made at the 3.6-m, where bright "AirCo" and "FloorCo", M1 P would 3. "Mirrar Seeing" and "Dome Seeing" patterns of flying shadows could be decrease sufficiently. This did not associated with both the slit area and the happen: Even after reducing major heat Mirror seeing is due to convection of air immediate neighbourhood over M1. This sources in the cage in 1994 (especially by above the reflective surface of a mirror could easily be checked measuring the TO ventilating the cage and the four when it is out of temperature equilibrium structure function in those two areas.

7 Top nng alf TO observers: The wind helps the seeing! • After reaching temperature equilibri­ um in the dome within 1.2° thanks to a particular P setting (always recorded) in the dome, the same result was obtained

Hub Dow pOlnl TO without wind ($ 2m/sec). secuo" Fans around Ml (18) • As soon as the primary mirror is T~2°, M 1 surlace eoollng system warmer than the outside by an 10 degradation due to M1 w.r.l. the seeing monitor 10 was identilied and detected: For T=2°, L'>(FWHM)=0,45" on average atthe 3.6-m.

5. New M1 CoolingNentilation System at the 3.6-m An air-handling unit (AHU) was in­ stalled in the telescope horse shoe. The cold air is presently se nt inside a circular tube located underneath the M1 cell, and from there through the cell to the mirror boltom thanks to a number 01 holes in the cell up and down. A second system is under construction, where the air is split before entering the cage. The second part will feed a circular "garden hose" located above the mirror on the side. This air will be cleaned and dried belore sweeping the mirror gently, day and nighl. P homoge­ neity will be obtained during the day thanks to 18 ventilators located above M1. Calculations show that surface P (5 ---:- Heal controller HoalOr cm in depth) can be changed by 1° in 3 Coollng hours, which would allow the outside P to COOLING be followed in most cases. It is intended to Fan SYSTEM (Air Hand· use only this new system plus the FloorCo ling Unit) during the next lew months in order to LOCATED ... IN HORSE . evaluate their impact on the seeing SHOE separately.

FIIler (oloctroslallc orHEPA) 6. Conclusion Figure 4: Mt eooling system presently implemented (sec/ion). Tests made during more than a year to measure the seeing with a dome in better thermal equilibrium were encouraging: Many times we crossed the fatelul When there are no heat sources or long (see refs. 3). That will be the subject of threshold of 1" with a regular instrument, time constant items in the telescope beam another paper, actually in preparation. reaching 0.8" on lour occasions, when neighbourhood, the overall seeing is not Nevertheless, Figure 3 shows the last in addition the primary mirror was cool. dominated by dome seeing, but rather by results obtained in March with a ther­ The next step is the constitution of a mirror seeing. mal environment which sometimes database, to record the seeing lor a Those pupil images, recorded with the approached perfection (L'>T", 0.9°). The maximum of meteorological and seeing autoguider and an out-of-focus beam, most signilicantlessons from those nights condition~. From those data and many were extremely useful in identifying (10 nights were allocated to seeing others recorded at the same time, we problems. Several patterns of Ilying measurements in period #54) are: will calculate a number of "10 indicators" shadows, present on 3.6-m pupil images, • Scientific runs: With the AirCo sys­ (=> image FWHM = I (parameter)), from never appear on NTT pupil images. Good tem alone, images got better for CO+ which seeing properties will be inferred. seeing was always associated with (0.3" improvement on average) while From there, we should know pretty weil vanishing 01 middle-scale Ilying shadow EFOSC1 10 was degraded (up to 0.5 how to tune up the ventilation/cooling patterns. Since this phenomenon is arcsec). Internal ventilation of EFOSC1 system in order to get the best possible dynamic, a picture unlortunately does not helped, but was not enough. EFOSC1 images. Then installing EFOSC2, whose say very much. is completely closed, CO+ completely pixel size is about hall that of EFOSC1 open. (0.61 "), at the 3.6-m Cassegrain locus, 4. EFOSC1, "Naked" CCD • When the dome was ventilated by the will allow us to really take advantage 01 the at the Cassegrain Focus wind ('" 6 m/sec), and when M1 and the good seeing. and Come-On-Plus air above M1 had the same P within 1.5°, the best images measured with a CCD Acknowledgements The results obtained during a year with were for a dome opened to the wind, EFOSC1 and especially with Come-On­ telescope at 20° from zenith: '" 0.8" This article summarises the work 01 Plus (CO+) will not be presented here FWHM. This fact is weil known to the a team. Thanks to the many discus-

8 sions with E. Swinnen and G. Ihle, we Special thanks go to C. Perrier who took 2. M. Faucherre, 1995, Internal ESO report: eventually managed to obtain a more the data with SHARP presented here. "3.6-m Seeing Improvement Project". accurate picture of the 3.6-m thermal 3. M. Faucherre, 1994-1995, Internal memo­ problem. References randae # 3.6-m/seeing-2, -11, -17 and -21. For the measurements I want to 1. Le Poole, 1990, final report: "Minimising the acknowledge: A. Gilliotte, T. Höög, M. man-made deterioration ofthe seeing at the M. Faucherre Maugis, A. Pizarro and G. Timmermann. 3.6-m telescope at La Silla". e-mail: [email protected]

CAT/CES NEWS L. PASQUINI, L. KAPER, ESO

Ouring the last week of March, a aresolution of 100,000 resulted in an the Long Camera starting in August 1995 new CCO was tested on the Coude actual resolving power of about 70,000. after the "idie" period of the CAT tele­ Echelle Spectrograph's Long Camera. The details are given in Table 2. scope. The Short Camera and CCO #9 This CCO, ESO#38, is a LORAULESSER According to the CCO detector group, the will be decommissioned and CCO #38 2688 x 512 thinned, backside illuminated degradation in resolution is expected offered to the observers requiring a device (pixel size 15 x 15 pm) with especially in the UV with backside resolving power up to R = 70,000. For anti-reflection coating. The quantum illlJminated devices. Oue to a fiefd-free programmes requiring higher resolution efficiency is about 80% throughout the region inside the device, photon-gen­ CCO #34 will be retained. A new version visible wavelength range (350-800 nm) erated electrons spread to adjacent of the CAT+CES Operating Manual with a peak value of 90% at 700 nm. The pixels, thus increasing the effective pixel containing the characteristics of the new values are better by afactorof 5 in the blue size. This effect is more pronounced in the configuration will also be distributed. The to 2 in the red than CCO#34 which is blue than in the red. high OE, the low read-out noise (8 e/ presently in use on the Long Camera (see Given the above results it was decided pixel), and the large size of CCO #38 Table 1). The high OE is obtained after not to offer CCO #38 to the ESO would be a significant improvement in flooding the CCO with intense UV light. In community at the start of the current the performance of the CAT/CES spec­ normal operations, it is expected that the period 55. For the moment, the Short trograph if a procedure for recovering the CCO will need to be UV flooded once Camera with CCO #9 and the Long expected spectral resolution can be every month. The new chip is mounted in Camera with CCO #34 are available. The developed. a continuous flow cryostat, with a hold stability of the UV-flooding of CCO #38 Some details of the characteristics of time of about one week. will be furthertested and a solution has to CCO #38 are provided in Tables 1 and 2. Efficiency tests were carried out which be found for the degradation in resolu­ ----- .------, ..._------~ confirmed the high sensitivity of the CCO. tion. Oue to the very high performance of We were, however, confronted with a this chip, we plan to ofter CCO #38 with L. Pasquini degradation in resolution at high resolving e-mail: [email protected] powers. Specifically, a slit setting to yield TABLE 2: CES LONG CAMERA + CCO #38 MEASUREO RESOLUTION VS NOMINAL RESOLUTION AT 4435 A TABLE 1: OVERALL CAT+CES LONG CAMERA EFFICIENCY IN PERCENT Nominal FWHM (Pixels) Measured Meas/Nominal ------~----- A(Angsträm) eeo #38 eeo #34 40,000 5.8 39,300 0.98

.~_ .._- - 50,000 4.9 46,500 0.93 3500 0.8 0.15 60,000 4.1 55,660 0.92 3589 2.3 0.47 70,000 3.75 60,800 0.87 4035 5.4 1.4 80,000 3.6 63,300 0.79 4435 6.9 1.1 90,000 3.35 68,000 0.76 5400 9.2 3.8 100,000 3.18 71,800 0.72 6450 10.4 5.2 110,000 3.0 76,000 0.69 8092 5.88 3.7 120,000 2.96 77,000 0.64 ----, •..... ------~.------

The FORS Focal Reducers for the VLT a Status Report

G. RUPPRECHT, ESO-Garching

Introduction a variety of observing modes in the visual 3. multi-object spectroscopy (MOS; up The FORS1 and FORS2 FOcal and near ultraviolet wavelength range, to 19 objects) Reducer/low dispersion Spectrographs namely 4. polarimetry (FORS1 only). are expected to be something like the 1. direct imaging (2 image scales) These modes can be combined e.g. to workhorses of the VLT since they will offer 2. low-dispersion grism spectroscopy allow imaging polarimetry or spectropo-

9 already been delivered to the eonsortium (by Halle). The standard set of grisms has also been designed and is presently being proeured. The broadband and interfer­ enee filters wh ich will be provided with FüRS are being defined in eollaboration with the Instrument Seienee Team. Sinee it beeame clearthat atmospherie dispersion would signifieantly degrade the image quality of FüRS, ESü deeided to install atmospherie dispersion eorree­ tors (ADCs) in unit teleseopes UT1 and UT3. A novel design ealled "Longitudinal ADC" eonsisting of two siliea prisms with variable distanee will be used. A paper giving full details is eurrently in prepara­ tion.

Mechanics

For reasons of eeonomy, instrument eomponents proeured externally are alwaysbought simultaneouslyfor FüRS1 and FüRS2. The Göttingen workshop is therefore presently busy with the ineoming inspeetions of the major meehanieal eomponents produeed by industry. So far, the eylindrieal housings for the instruments have been reeeived Figure 1: The FORS Wollaston prism; note the enormous size! (Fig. 2) as weil as the foeusing units made of cast aluminium and the meehanieal eomponents (spindies, linear bearings) for the multi-objeet speetroseopy units; larimetry. In addition, a mode offering redueer) are presently being polished by also, a signifieant part of the manufaetur­ higher dispersion (possibly up to R ­ FISBA. Delivery is expeeted later this ing is done in Göttingen. 5000) is envisaged for FüRS2 whieh will year. The opties forthe polarimetrie mode, therefore not be equipped with polarimet­ which eonsist of a Wollaston prism of 132 Electronics and Software rie opties. An overview of the FüRS mm (!) free diameter (Fig. 1) and two 3 x projeet ean be found in [1]. A more 3 mosaies with 135 mm free diameter of This part of the projeet is performed in eomprehensive deseription of the expeet­ superaehromatie retarder plates, have Munieh. Most eleetromeehanieal units ed instrument performance is given in [2J and [3] and more teehnieal deseriptions in [4] and [5].

Status of the Project

Work is progressing weil at the three institutes (Landessternwarte Heidelberg, Universitäts-Sternwarte Göttingen, Uni­ versitäts-Sternwarte München) eollabo­ rating in a eonsortium. Already in April 1992 the Preliminary Design Review (PDR) was held, and in February 1994the Final Design Review (FDR) followed for the instrument meehanies, eleetronies, the assembly, integration and test proeedures, for handling and mainte­ nanee aspeets and for safety and management issues. No serious prob­ lems were identified, so immediately afterwards the eonsortium began to transform the approved design into hardware.

Optics

In Heidelberg, the design of the Figure 2: Incoming inspection of the FORSt and FORS2 top sections; they will house the multi­ imaging opties has been finished and the object spectroscopy units. At right are the tubes of the collimators for FORSt needed to change lenses (FüRS is an all-dioptrie foeal the image scale.

10 (motors and encoders) have been nance procedures and to reduce the of 1998, according to the current VLT procured. diversityofancillary equipment atthe VLT, planning. FORS2 is then scheduled for an In September 1994 the FDR for some the carriage is therefore being designed installation on UT3 in the year 2000. of the instrument control software to accommodate both; two copies will be modules was held and coding started procured in 1995. References immediately afterwards. Data reduction software for FORS [1]1. Appenzeller, G. Rupprecht, "FORS - The specific instrument modes (mainly MOS Detector Focal Reducer lor the VLT" ESO Messen­ and polarimetry) is being implemented in ger 67, p. 18 (March 1992). [2] G. Rupprecht, "FORS" in: Instruments for Heidelberg. The first finished context One of the most crucial components of the VLT, ESO Instrumentation Department, (MOS) will be included in the 95NOV an instrument is its detector. For FORS A. Moorwood, Editor, p. 9 (1994). release of MIDAS. SITe (formerly Tektronix) 2048 x 2048 [3] W. Seifert, W. Mitsch, H. Nicklas, G. CCDs with 24 11m pixels were selected; Rupprecht, "FORS: A Workhorse Instru­ ment for the ESO VLT," in: Instrumentation the procurement of the CCDs including all Auxiliary Devices in Astronomy VIII, David. L. Crawford, Eric peripherals (dewar, controller) is being R. Crain, Editors, Proc. SPIE 2198, p. 213 done by ESO. So far we have received (1994). Several auxiliary devices are under one CCD of grade 1which is now awaiting [4] W. Mitsch, G. Rupprecht, W. Seilert, H. construction for FORS. The most full characterisation by ESO's detector Nicklas, S. Kiesewetter, "Versatile multi­ object spectroscopy with FORS at the ESO important ones for the construction and laboratory. test phase are the star simulator for the Very Large Telescope," in: Instrumentation in Astronomy VIII, David. L. Crawford, Eric optics tests, which is partly finished, and R. Crain, Editors, Proc. SPIE 2198, p. 317 the telescope simulator to be used mainly Future Planning (1994). for the f1exure tests of the integrated [5] H. Böhnhardt, S. Möhler, H.-J. Hess, S. instruments; this one is under construc­ Activities scheduled to happen in the Kiesewetter, H. Nicklas, "Design Bench­ tion. near future include tests of the imaging marks of the FORS Instrument for the ESO VLT," in: Scientific and Engineering Fron­ Another important device will be the optics and performance tests of the first tiers for 8-10m Telescopes, M. Iye and T. transport carriage. It turned out that the MOS unit, both from a mechanical and Nishimura, Editors (1995). requirements for handling the FORS control system point of view. instruments at the telescope are very The long-term planning foresees the similar to those of the Cassegrain integration of FORS1 for 1996, system G. Rupprecht, ESO-Garching adapters/rotators, e.g. weight to be tests for 1996 and 1997, transport to the e-mail: [email protected] transported, lifting height and mounting VLT Observatory in 1998 with an URL: http://vll.usm.uni-muenchen.de:8002/ accuracy. In order to simplify mainte- installation date on UT1 in the last quarter home.html

UVES (UV-Visual Echelle Spectrograph) for the VLT - a Status Report

H. DEKKER, ESO Instrumentation Division

Introduction objectives and expected performance of Trieste. S. D'Odorico is the Instrument the upgraded instrument are given in Scientist. UVES is a two-arm crossdispersed Tables 1 and 2 and in [4]. echelle spectrograph covering the wave­ UVES is being designed and built in­ Status of the Project length range 0.3-0.5 11m (blue) and house. The instrument control and data 0.42-1.1 11m (red), with a 2-pixel reduction software is being developed in Following the Preliminary Design Re­ resolution of up to 90,000 and 120,000, collaboration with the Observatory of view in October 1993, UVES is now in the respectively. It will be mounted at the Nasmyth of UT2. Project kick-off for UVES was in spring 1992 with a plan [1] calling for two TABLE 1. UVES SCIENCE OBJECTIVES ,------_.. _----- identical instruments, to go on UT2 and • Structure, physical conditions and abundances of interstellar and intergalactic gas at UT3, with aresolution of up to 70,000. An early epochs from the absorption spectra of high redshift QSO's overview is given in [2]. In response to • Kinematics of gas and stars in galactic nuclei discussions on aredistribution of • Kinematics and mass distributions of star clusters instruments at the foei of the VLT, it was decided in spring 1994 to build a single • Composition, kinematics and physical conditions of the interstellar medium in the galaxy and in nearby systems instrument with increased spectral resolution and versatility (by adding an • Chemical composition and atmospheric models 01 galactic and extragalactic stars Atmospheric Dispersion Compensator, • Substellar companions of nearby stars (high-precision radial velocity studies over long time scales) an lodine absorption cell, a depolariser and exposuremeters) [3]. The science • Stellar oscillations

11 TABLE 2. UVES OBSERVING CAPABILlTfES AND PERFORMANCE and a minumum OE of 75% at 600 nm (previously 60%). These devices would Blue Red be weil suited to the red arm. Another contract with the University of Arizona Wavelength range 300-500 nm 420-1100 nm (Lesser) is in place that covers the need Resolution-slit produet 40,000 40,000 for a UV-blue sensitive 2K chip needed for Max. resolution 90,000 120,000 the blue arm; several devices were recently supplied by this source with a OE Deteetion efficieney 9% at 400 nm 10% at 600 nm of over 70% in the UV. ESO is closely Limiting magnitude 18 (R =50,000) 20 (R =45,000) following developments at SITe, MIT/LL (3 h. exp. time, S/N = 10) in U 18.5 (R =90,000 and EEV; all potential sources of 2K x 2K in V Camera dioptrie F/1.8, 70 J..lml" dioptrie F/2.5, 97 ~lml" or 2K x 4K chips with 13.5 or 15 flm pixel field 43.5 mm diam. field 87 mm diam. size. The prototype ESO CCD Array Control Baseline CCD and pixel seale 2048 x 2048, 4096 x 2048 15 J..lm pixels 15 J..lm pixels, Electronics (ACE) intended for use with (.215 "/pix) 2 x 1 mosaic (.155"/pix) FORS and UVES has been successfully Eehelle 41 .59 g/mm, R4 mosaie 31.6 g/mm, R4 mosaic tested at the NTT in early 1995. A prototype continuous-flow cryostat has Crossdispersers CD1: 1200g/mm, CD3: 600 g/mm, been built and tested in the CCD detector Ab 380 nm Ab 550 nm CD2: 600 g/mm, CD4: 316 g/mm laboratory and field experience with this Ab 380 nm Ab 750 nm system - containing one of the Lesser J",JJframe (typ) 700 A in 20 orders 1000 A in 18 orders chips - is being obtained at the CES in La Silla. Order separation (typ.) > 15" H 70 pixels > 15" H 100 pixels - ----_..------

Schedule detailed design phase. The Critical De­ Electronics and software - The first sign Review is planned for November batch of UVES functions is under test. Important milestones for 1995 are the 1995. Detailed status is as foliows: Electrical design for components that delivery of the red echelle and Thomson Gratings and Optics - In view of the were added in the UVES upgrade (e.g. CCDs, placing of the optics contract and new technology involved, a contract for exposure meter) is in progress. An the Critical Design Review at the end of the large (84 x21 cm) echelles was placed updated Electronic Design Report, the year. 1996 and 1997 will be devoted to al ready in December 1993 with Milton addressing the new overall system completion of hardware, software coding Roy (USA). They are monolithic mosaics: architecture including the Scientific CCDs and lower-Ievel testing. Extensive sys­ replicas on a single substrate of a 2 x 1 is planned for the 2nd quarter of 1995. tem testing, including calibration and mosaic of submasters. The red echelle Various software specifications that data reduction procedures (using the has recently been completed and is between them address all aspects of Sun) will take place in 1998. Commis­ awaiting acceptance test. Manufacturer's operating UVES and reducing data sioning as instrument 1 on UT2 is preliminary test results indicate that this is obtained with it have been released. planned to commence in July 1999. one of the best gratings ever made (Table (Software Requirements, Software Func­ 3). Further testing, including the support tional Specifications, Data Reduction References system, is planned at ESO. The blue Software Requirements). A draft user's echelle mosaic will be delivered in 1996. manual and a prototype Graphical User [1] UVES Design and Implementation Plan, ESO/STC-130. Available fram the VLT Ofters from industry for detail design Interface are planned to be produced this Archive under No. VLT-PLA-ESO-13200­ and manufacturing of cameras and preslit year. 230. optics according to the ESO predesigns Detector systems - The parameters [2] H. Dekker and S. D'Odorieo , "UVES, the have been received; contract negotia­ of the UVES cameras are matched to UV-Visual Echelle Speetrograph for the 2 VLT", The Messenger70,p. 13(Dee. 1992) tions are in preparation. Deliveries are 2048 , 15 flm pixel size chips (blue: single 2 [3] Upgraded UVES Design and Implementa­ expected in 1996 (preslitoptics) and 1997 chip; red: mosaic up to 4096 ). The CCD tion Plan, ESO/STC 151. Available from the (cameras). detector systems for UVES are being VLT Archive under No. VLT-PLA-ESO­ Mechanics - Prototype drive units developed by the newly created detector 13200-0596. have been produced and tested. A first group within the Instrumentation Division. [4] H. Dekker, "UVES" in: Instruments for the batch of 9 motorised functions (slides, ESO has placed a contract for the VLT, ESO Instrumentation Department, A. filter wheels) has been delivered and development of thin, 2K x 2K, 15 flm Moorwood, editor, p. 19 (May 1994) electromechanically tested. Slit, derota­ buttable CCDs with Thomson CSF in tor and crossdisperser units are in detail 1992. Recent negotiations with Thomson H. Dekker design and/or manufacturing. By the end led to agreement on delivery within 1995 e-mail: [email protected] of 1996, most fine mechanical and/or motorised units will have been delivered. The detailed design of the table and support structures is almost complete. TABLE 3. PRELIMINARY MEASUREMENT RESULTS OF UVES RED ECHELLE (31.6 g/mm) The earlier concept of the enclosure is Speetral resolution 3.0 mA at 632.8 nm being compared with a simpler concept R =2,100,000 with possibly lower cost and better thermal and light rejection performance. Angular resolution in the direetion of the slit better than .1 arcsec on the sky Also the handling tools are under review. Blaze angle 75.07° A full-scale wooden model of UVES has Ghosts < .008% been made to develop handling concepts, Absolute blaze effieieney (ineluding dead spaee) 72.4% at 550 nm study cabling routing, etc.

12 Result of The Phase A Study for the VLT Mid-Infrared Instrument: VISIR

Po. LAGAGE, Y RIO, O. OUBREUIL, CEAIOSMIOAPNIAIService d'Astrophysique, CE Saclay, Gif-sur- Yvette, France TH. OE GRAAUvv, SRON, Groningen, the Netherlands J. W PEL, T SHOEMAKER, Kapteyn Observatory, Roden, the Netherlands

In March 1993, a contract for a Phase ratio can be reached 25 times faster. "warm" dust (400 K- 140 K). Indeed, a A study concerning the mid-I R instrument When taking into account improvements black body at400 K has its peakofthermal to be mounted on the VLT unit2 (Refs. [4] in detector performances, telescope emission at 7.5 pm (the beginning of the and [5]) was signed between ESO and emissivity and image quality, "dome" 10 ~lm atmospheric window), while a DAPNIAICE Saclay, with as partners seeing, etc., we can expect a gain in black body at140 K has its emission peak SRON at Groningen and the Kapteyn sensitivity by a factor 50 compared to the at 28 pm (about the end of the 20 pm Observatory at Roden. The study is now actual TIMMI sensitivity (Refs. [2] and atmospheric window). Dust as cold as 50 finished and the main conclusions are [3]). Although the VLT mid-IR instrument K should still be detectable at20 pm. The outlined here'. Various observing modes will be much more sensitive than IRAS warm dust is an important component of are foreseen both in the N and Q (band 1 and 2), it will not compete with the the Universe, as shown by the IRAS atmospheric windows (10 and 20 pm): instruments on board the cryogenic satellite. The study of this dust is related • imaging over a field up to about 1 cooled Infrared Space Observatory to actual astrophysical problems dealing arcmin with various magnifications « 0.08 (ISO), except for the high spectral with star formation, planet formation, arcsec per pixel to 0.3 arcsec per pixel), resolution mode. However, ground­ stellar evolution, AGN unified scheme, • long (> 30 arcsec)-slit spectroscopy based mid-IR instruments on a large etc. and concerns various astronomical with various spectral resolutions (350, telescope allows for a diHraction-limited objects ranging from comets to quasars. 2500,25,000 at 10 pm and 1250, 12,500 angular resolution of 0.3 arcsec at 10 pm The observing modes related to these at20 pm), (1.22 /JD), inaccessible with the small programmes are: • limited polarimetry. ISO telescope (60 cm). Thus, ground­ • diHraction-limited imaging both at 10 The instrument will be an excellenttool based and space observations are com­ and 20 pm to locate the dust, measure its for ISO follow-up studies. Even if its sen­ plementary and we can foresee a lot of temperature, sitivity will not compete with ISO, except ISO follow-up observations. This per­ • low spectral resolution (R = /Jot.. = for the high spectral resolution mode, its spective leads us to stress the need tor 400) to constrain the composition of the high angular resolution (diHraction limit­ such an instrument soon atter 150. grains (for example looking for the silicate ed) will make it very allractive. To be more explicit on the observing dust features at 10 and 18 pm). modes, we now discuss two types of ob­ • polarisation to learn about grain size. 1. Scientific Needs servations: observations of dust and gas. 1.2 Gas Studies Designing an instrument is not an easy 1. 1 Dust Studies task because many parameters (some­ A wealth of information about the times not completely rational!) have to be The two mid-I R atmospheric windows gaseous component in a large number of taken into consideration and trade oH has are key domains to study relatively various objects can also be provided by to be made between scientific desidera­ ta, technical and operational constraints, costlimitation, etc. The easiest part is to get the scientific desiderata. If a large enough community is consulted, the answer is simple: al/. More seriously, generaltendencies of the observing programmes and associated observing modes can be c1eared Up2. First, we recallthatthe use of amid-IR instrument on a large telescope will lead to a dramatic improvement in sensitivity. Indeed, the size diHerence between the VLT and a 3.6-m telescope leads immedi­ ately to an increase in sensitivity by a factor 5 for point source observations; this means thatthe same signal-to-noise

IThe eomplele report is available upon request al [email protected]. 2Speeifie programmes ean be lound in the Phase Areport, in lhe proeeedings 01 lhe Freneh meeling "L'exploilalion astrophysique des lenelres 10 et 20 pm avee le VLT", Oelober 1993, D. AIIoin and P.O. Lagage Eds., and in the proeeedings 01 the ESO workshop aboul "Science wilh the VLT", June 1994, Danziger and Walsh Eds. Figure 1: Opticallayout of VISIR.

13 observations in the thermal infrared. Indeed, many molecular, atomicand ionic spectral lines are present in this wavelength range. Of special interest are several key interstellar and circumstellar symmetrical molecules (H 2 , CH4 , CO2 , etc.) which have to be detected through observations of the (vibration-)rotation lines in the infrared. The linewidth varies from a few km/s in stars to a few hundreds of km/s in active galaxies. Several spectral resolutions ranging from a few thousands to about 100,000 are needed to cover the various cases. Here again, observations both at 10 and 20 11m are interesting. For example, from

the ratio of the H2 1ines intensity at 12 and 17 11m, the gas temperature can be derived. We have set two regimes, one for medium-resolution (2500 at 10 11m) and one for high spectral resolution, limited to 25,000 (at 10 11m) to avoid technical configurations which appear too risky. Figure 2: Optical arrangement ofthe camera and the low andmedium spectral resolution modes. In the imaging mode, the light goes first through a collimator lens which images the telescope 2. Instrument Design secondary mirror onto a cold pupil stop. After passing the filter, the light is re-imaged on the detector array with various magnifications. thanks to one ofthe lenses ofthe second lens wheel. The optical arrangement of VISIR, an This wheel also holds a lens followed by a folding mirror to image the focal plane onto the s/it unit instrument which fulfills the previous with a magnification of0.8. After the slit, a folding mirror, M2, selects one of the two spectrograph scientific requirements, is shown in arms. Then the light enters a 3-mirror collimator of the ISAAC type, which produces a collimated Figure 1. VISI R stands for VLT (mager beam of55 mm on the selected grating. The selection and scan of the gratings are based on the Spectrograph in the IR. It is made of two mechanisms made for the ISO/SWS instrument. After reflection/dispersion on the gratings the light goes back throughout the 3 mirrors, which are now used to image the spectrum on the array, sub-units: an imager and a spectrograph. via the folding mirrorM6. The magnification ofthe system is 1. (Note that in Figure 1, M6has been The spectrograph has two arms, one for inclined differently so that the spectrograph detector is near the imager detector for cryogenics the low and medium spectral resolution, reasons; a filter wheel cooled at 15 K has been included in front ofthe spectrograph array to limit the other for the high spectral resolution. the internal photon background emitted by the spectrograph optics cooled at 45 K.) The whole optical bench is enclosed in a cryogenic/vacuum vessel to prevent in­ ternal background; (a black body at room temperature has its peak emission at ... 10 11m). The vessel is a cylinder, 1.1 m long and 1.7 m in diameter. The total instrument weight is 1.5 tons. Closed­ cycle coolers are foreseen to maintain the optical bench at the required tempera­ ture: 45 K for most of it and 15 K for the parts nearthe detector. The detectoritself has to be cooled down to about 5 K (to prevent dark current). The same arraywill cover both the N and Q bands. The detectorarray used forthe imagerand the spectrograph will be different in order to be adapted to the quite different background flux received when observ­ ing in broad-band or at high spectral resolution.

2.1/mager The imager design is c1assical and based on lenses, in order to easily implement various magnifications. In­ deed, given the state of the art in detector complexity (256 x 256), it is not possible to have both a large field and over­ sampling of the diffraction pattern, while Figure 3: Opticalarrangementin the high-resolution mode. The beam goes through the imageras one or the other is required according to in the lowandmedium spectroscopic mode. The folding mirrorM2 has been f1ipped by90 degrees the programmes. For example, pro­ to deviate the lightinto the high spectralresolution arm, also made ofa 3-mirrorcollimator/collector of the ISAAC type; the collimator beam is now 110 mm. The grating unit is made of two nearly grammes aiming at detecting faint similar echelles, mounted back to back. A second filter wheel has been included in the imager to extensions around stars will prefer over­ sort out the various orders. The mirror M6 has also been f1ipped by 90 degrees to send the sampling of the airy pattern (PFoV < 0.1 spectrum on the same detector array as in the low and medium spectroscopic modes.

14 arcsec at 10 IJm), even if the field will be eventually found to be on the safer cy reasons, we plan to use the "duo­ limited to less than 25 arcsec. On the side. echelle" concept: two nearly similar other hand, observations of objects quite When working at 10 IJm, we are in a echelles, mounted back to back, where extended or survey-type observations world where diffraction is around. To avoid the orders of one fit between the orders of will require a larger field. With a simple diffraction at the slit, we have considered the other echelle. It is worth mentioning two lenses design (see Fig. 2), it was a 2 A. F-ratio entrance slit. Even with such the size of the gratings: 35 cm! The possible to implement a field up to 80 a slit width, the optical elements of the efficiency of the optics (including order arcsec with diffraction-limited optical spectrograph have to be oversized to sorting filters) should be around 40% for quality. avoid diffraction losses. It is also better to the low and medium spectral resolution In ourtwo-Ienses design, the first lens have the cold pupil stop in front of the slit and 30% for the high spectral resolution. images the pupil of the secondary mirror because, behind it, the pupil is fuzzy. In The detector array used for the of the telescope on a cold stop to prevent Figure 1, thecold stopofthe imagerisalso spectrograph will be optimised for low extra external background (as usual in used for the spectrograph. In a previous background conditions. Indeed the flux ground-based IR instruments). The lens design, a separate cold stop/re-imaging received by the spectrograph can be up to also acts as a collimator, so that the filter unit based on two off-axis paraboloids 105 fainter than in the imager. It is more in wheel located near the cold stop is in a and two folding mirrors was considered. the range of the f1ux received when parallel beam. Such a configuration Both options are open. Note that the use imaging from space, and detector arrays relaxes the tolerances on the spatial of a re-imaging unit in front of the have been optimised forthese conditions. homogeneity of the filters and leads to a spectrograph has additional advantages; The pixel pitch of these detectors range focalisation wh ich is independent on the it provides the possibility to move the from 30 to 50 IJm. In the spectrum focal filter whatever its thickness iso But the spectrograph entrance slit to a more plane, 50 IJm represent 0.115 arcsec on drawback of this configuration is a slight convenient location; it allows to reduce the sky, wh ich sampies correctly the airy change in the wavelength transmission the F-ratio (from 13.6 to 11) to keep the pattern at 10 IJm. To avoid additional of the filter according to the position in the spectrograph dimensions within limits; it complexity, we have not considered, at field (different angle of incidence on the allows to inject the parallel beam of the the present stage, a re-imaging system filter). To avoid prohibitive shifts in internal (wavelength) calibration source, after the spectrograph. Note that the good wavelength for narrow-band filters (R = which will be quite similar to the source image quality of the ISAAC 3 mirror 50), the cold stop has been fixed to 15 used for the ISO/SWS instrument. system allows forthe implementation of a mm. There is actually no good optical The spectrograph is based on the 3­ detector up to 1024 x 1024 pixels. material that covers both the N and 0 mirror collimator/collector design devel­ The mechanical tolerances on the bands. We will use Germanium for 10 IJm oped by ESO for the ISAAC instrument, spectrograph are quite tight for a and CdTe for 20 IJm; the two collimators one of the near IR instruments to be cryogenic cooled instrument. A finite­ will be mounted on a wheel. A second mounted on the VLT unit 1 (Moorwood et element mechanical study was made for wheel will hold the lenses which re­ al., 1993). The light entering the slit goes the previous optical design studied in images the focal plane onto the detector through the three mirrors, wh ich produce detail. The tolerances on the present with various magnifications (0.08 arcsec, a collimated beam on the gratings; once design are tighter by a factor 3; (for 0.16 arcsec, 0.3 arcsec). The antireflec­ diffracted/reflected by the gratings, the example, the tilt between the 3-mirror tive coatings on the optical materials beam goes back through the 3 mirrors collimator and the detector should vary by considered are very good (a few per cent which now act as a collector to image less than 1 arcsec when moving the of reflection), so that the efficiency of the the spectrum. The 3-mirror ISAAC colli­ telescope). We are in the range of the optics will mainly be determined by the mator/collector has many advantages in ISAAC tolerances, but we are not at the filter transmission. By an appropriate terms of compactness, optical quality, same focus (Cassegrain instead of design, the ghosts resulting from light straylight rejection, grating efficiency Nasmyth), so that we have to wait for very reflection on the lenses can be made (grating in Littrow configuration), which detailed finite element calculations to fully negligible. overcomes the relatively large number of address the mechanical question. Limited polarimetry capability will be mirror reflections. provided by 3 analysers in the filter wheel. The compactness of the three-mirror 3. Sensitivity Several firms (LETI/L1R in France ISAAC collimator led us to consider two and now RockweIl and SBRC in the US) arms forthe spectrograph, one forthe low The goal is to achieve the theoretical have developed detectors optimised and medium resolution with a collimator expectations based on background noise for broad-band observations from the beam of 55 mm, and the other for the high limited performances and a telescope ground. In these detectors, the storage resolution with a collimator beam of 110 emissivity of 10%. The sensitivity of the capacity of a pixel has been pushed to its mm (see Figs. 2 and 3). Four gratings are imager in broad-band observations of maximum (> 3 107 electrons) in order to planned for the low- and medium­ point sources should tend towards 1 mJy "absorb" the huge photon background resolution modes; for efficiency reasons, 10 cr 1 hour at 10 IJm (or magnitude 11.5) generated by the telescope and the two gratings are needed for a band (0 or and 10 mJy 10 cr 1hourforthe 0 band. The 2 atmosphere (1500 Jy/arcsec ), without N); but the same gratings are used for the theoretical sensitivity of the spectrograph being saturated. o and N sub-bands (1 st order for 0, 2nd is given in Table 1: order for N). The high-resolution mode is The high spectral resolution is driving 2.2 Spectrograph also based on gratings, but used at high the specifications in many domains: The design of the spectrograph and orders (echelle mode). Again for efficien- detector noise, internal background, especially its high-resolution mode was the sUbject of many debates. To the 3 optical pre-designs internally made at TAßLE 1. SPECTROGRAPH SENSITIVITY (Signal/noise in 1000 s tor a 1 Jy point source) ESO (Ref. [1] and references therein), Od we added a couple of alternative de­ signs. Hereafter, we will only describe 1-2}1~~~f::-=t_-~-r--- the final one (we should say the last -O:i - one). The slit option with gratings was J

15 mechanical tolerances, . The per­ too numerous to be quoted here. We also Ed. Moorwood A.: 1993, "IR Instrumenta­ formances of this mode are the most would like to thank the ESO staff and tion tor the ESO VLT" in Infrared oetectors and Instrumentation, SPIE volume 1946, p. especially A. Moorwood, B. Oelabre, H. difficult to achieve. 461-471, Cormer B. Ed. Käufl, J.-L. Uzon a l'Aliemand for their [3] Käutl H.-U., Jouan R., Lagage P.O., Masse 4. Perspectives valuable comments all along the study. P., Mestreau P., Tarrius A.: "TIMMI at the 3.6-m Telescope". The Messenger70, 67. VISI Ris an ambitious instrumentwhich References [4] Käutl H.-U.: "Phase-A Study Launched tor will require a total work of 100 man years the 10/20-pm Camera/Spectrometer tor [1) Käutl H. and Delabre B.: 1994, "Improved ESO's VLT". The Messenger72, 44. with some hard time. Design and Prototyping tor a 10/20 pm [5] Käutl H.-U.: "Ground-Based Astronomy in Rendez-vous in five years to report on Camera/Spectrograph tor ESO's VLT" in the 10 and 20 ~lm Atmospheric Windows at the first telescope tests of VISIR, . Instrumentation in Astronomy VIII, SPIE ESO - Scientitic Potential at Present and in hopefully. volume2198, p. 1036-1047, Crawtord D.R. the Future". The Messenger73, 8. Acknowledgements: We wish to and Craine E.R. Eds. thank the various colleagues who have [2] Lagage P.O. et al.: 1993, "TIMMI: a 10 pm camera tor the ESO 3.6-m telescope" in contributed to the phase A study, either Infrared oetectors and Instrumentation, P.O. Lagage technically or scientifically and which are SPIE volume 1946, p. 655-666, Cormer B. e-mail address:[email protected]

~ ~BIT " & I"~I,I I PIXELS;

With this periodically compiled collection of short notes, the NTT Team intends to keep the community informed about changes in performance, configuration, and operation of the NTT and its subsystems.

First NTT Team Member To leave This is the first big step towards the of work with the NTT Team at La Silla operational part of Phase II and beyond of should be the optimal practical and The first departure of members from NTT UP which foresees that the NTT will conceptual preparation for young astron­ their group often marks the transition to a be re-commissioned in service mode, omers with strong interest in observa­ new phase. This is also true for the NTT much the same as is planned for the tional work for the VLT era. Currently, we Team which started to come into exist­ commissioning and operation of the VLT. have an opening for arecent PhO ence about one year and a half ago. Since Since this is a very important and recipient. If you are interested in a then, a fair number of improvements complex type of work, for which there is challenging job which offers an unusual could be reported, and the operation of only limited experience available, Gabriel diversity of experiences and research the NTT has stabilised considerably. and Roberto are joining us early in order opportunities at a major observatory, Thanks to his astronomical expertise, to insure the proper commissioning ofthis consult the vacancy notice in the An­ Edmond Giraud has had his share in this operating mode. nouncements section of this issue of The progress. When Edmond heard about the An important aspect of this early start Messenger. NTT Upgrade Plan (The Messenger 75, is that the training of the instrument 1), he spontaneously offered to spend a operators will not be limited to ESO staff. year on leave of absence with the NTT Between now and the 'Big Bang' on April Field Test 01 Work Component Team at La Silla. (Our colleague Miguel 1, 1996, Gabriel and Roberto will increas­ No.5 Albrecht has suggested the handy acro­ ingly assist Visiting Astronomers in the nym "NTT UP"forthe NTT Upgrade Plan, operation of EMMI and SUSI. We hereby Ouring February 18-21, the latest test which henceforth we shall gladly use). hope to achieve two goals: (a) The know­ of another component of the VLT-like That year has meanwhile become 14 how transfer will also take place directly control software as defined in NTT UP months long, but now it is time for him to from the community to the instrument was successfully completed. It consisted return to his home institute, the Observa­ operators; (b) Visiting Astronomers can of two parts, (i) the control of the hydraulic toire de Marseille. convince themselves that also in service system and (ii) the selection of video This is a good example of how ESO's mode their programmes will be in 'good signals such as from guide probe or slit role as a service organisation can be hands'. (The scientific responsibility and viewing cameras and their distribution to strengthened by the active support of the supervision will always rest with an the requested local or remote destina­ astronomical community. We thank Ed­ astronomer.) tion. Apart from these specific applica­ mond for his willingness to help during a We wish Roberto and Gabriel a tions, the scope of the tests also included critical initial phase and wish him all the successful start with their demanding checking the VLT LCU Common (LCC) best. work. and Central Control Software (CCS). Once again, no real problems were encountered during the installation so Instrument Operators to Join the Postdoctoral Fellowship Availa­ that most of the night time could be used Team ble at La Silla for further optomechanical tests and other work with the telescope. As for all On May 16 and June 1, Gabriel Martin Given the proto-typical character of othertests of the new control system, the and Roberto Aviles, respectively, will take the NTT hard- and software and the old control software was fully restored up their duties as instrument operators. operations model for the VLT, a few years after completion of the tests.

16 R4 Grating for EMMI the problems which are now occasionally Rotator Bearings Now Offered experienced with accessing this cata­ logue in Garching will no longer occur. The ball bearings of the instrument For a scaled-down prototype of the R4 rotators in the two Nasmyth stations are grating to be used in UVES (H. Dekker VME CCD Controllers operated under conditions which are very and S. D'Odorico, The Messenger 70, different from normal applications of such p.13), the commissioning could now be Visiting Astronomers Palle Möller and devices: the rotator bearings move con­ completed. A year earlier, this had been Steve Warren reported a strange prob­ tinuously, but hardly ever turn through made impossible by a problem with the lem which they noticed in data obtained angles of 360 0 or more, run at very low calibration of the slit width which was with SUSI and CCD #25. In exposures speed and regularly change direction of noticed only after the completion of the with low light levels, the noise exceeded motion. During the zero passages of the observations so that the actual resolving the value expected for Poisson statistics speed, the effects of the transition power could not be assessed. This time, by up to 40%. In images with very uniform between static and dynamic friction are the nominal slit-resolving power product signal level, a weak chessboard-like aggravated when, owing to the special of 70,000 arcsec (corresponding to 2­ pattern could be seen. In a remarkably operating conditions, the balls have pixel sampling with the F5.3 camera ofthe quick and concerted action of the La Silla settled at the bottom of the bearing. red arm of EMMI, cf. Dekker et al., The and Garching branches of the newly­ There are indications that rotating the Messenger76, p.16) could be confirmed formed CCD group, the problem could be bearing through 360 0 wh ich rotates any Over the full spectral format. The traced back to cross talk involving the bus cluster of balls from the bottom of the combination of high spectral resolution and analogue-to-digital converter of the bearing to its top, can keep the incidences and a wavelength coverage of 250 nm in VME controller. When the output of the AI of increased friction at alow level.ltshould the red and 450 nm in the 'blue' make this D converter changed strongly, i.e. near also be noted that these events have mode of EMMI a rather unique research powers of 2 where many bits flip, this almost never affected the observations. tool. As of Period 56 (starting October 1, effect was particularly noticeable. This However, on side A (IRSPEC, SUSI), 1995), it will be offered to the community. explanation predicted that the same additional problems have been noticed for However, in order to limit the additional problem would occur also for CCD #31 in quite some time which are not covered by operational overhead, it will be scheduled the blue arm of EMMI, since it is operated the simple model described above. In only if nights can be combined into one or with the same type of controller. In fact, particular during beam switching for more blocks of sufficient length. careful analysis confirmed this expecta­ IRSPEC observations, the bearing got tion, although to a lesser extent than in completely stuck a couple of times. After SUSI. By changing the timing of the the electronics problems with the rotator Closed-Loop Operation of Active readout and A/D conversion process the control had been eliminated (The Mes­ Optics System symptoms could be fully suppressed. senger 79, p.1 0), this problem accounted for roughly 0.7 of the 1.8% of observing Considerable progress has been time lost because of technical failures in made towards making the New Tech­ Image Quality Monitoring the first quarter of 1995. nology Telescope the optically fully active Programme and Enclosure To investigate this behaviour more telescope as it was originally conceived. Operations Model thoroughly, it was decided to open the On an experimental basis, the parallel bearing (including the re-installation, this mode of the image analysis has, with We gratefully acknowledge the efforts takes 3 days). Some small cracks were much help from Lothar Noethe and by Wolfgang Eckert and Juan Carlos noticed in the bearing races. The evalua­ additional advice from Krister Wiren­ Pifieda to manufacture and install the tion of this finding has just started. Once strand, been put into operation. This mount for an additional CCD camera. the bearing was open, this opportunity means that 80% of the Iightfrom the guide Attached to guide probe 2 of side B was also used to replace the balls and star is diverted to the image analyser (EMMI), it will be used to monitor the install a different type of spacers which which can thus be operated parallel to the FWHM of the images delivered by the separate the balls. In the first week of scientific exposures. The results can be telescope. This approach was chosen intensive usage of SUSI after this inter­ accepted by the observer during or after because the image size does not come vention, notone overcurrentwas recorded. the exposure, with the former option out of the image analysis nor can it be We thank Gerardo Ihle and the currently not being advisable for direct readily extracted from all scientific data, Mechanics Group forthe effective support imaging. e.g., spectra. We thank Lothar Noethe provided on relatively short notice. The So far, image analyses was possible and Francis Franza for their help with the above intervention was only possible only at the centre of the field, thus procurement of the CCD camera and the because the same group had upgraded excluding the simultaneous usage of the optics, respectively. the crane to the capacity needed for the scientific instrument. As a result, even In Garching, Volker Bäumer started to safe handling of the adapter. during nights of excellent seeing, most work on his PhD thesis which aims at The setting up of a test stand of aspare observers did not use the active optics establishing the link between model NTT bearing and motor in Garching was system as often as would have been calculations for the wind f10w in the ND nearly completed. It will be used in the optimal. Therefore, the recent step im­ enclosure, measurements of wind speed frame work of a research project at the proves both the operating efficiency and and direction, temperatures, and the Technical University in Munich to better the practical optical performance of the seeing. The mechanics workshop at La understand the nature of the problem and NTT. Since the image analysis is more Silla has provided a platform for a mobile to try out possible solutions. This was demanding on the brightness of the guide anemometer. Visiting astronomers will done by F. Franza, M. Ziebell, and B. star than the autoguider alone, tools have see it in operation mainly between the Gustafsson. We are grateful to them for also been developed to automatically opening of the dome in the afternoon and their continued support in analysing the select the brightest suitable guide star the beginning of the observations. bearing problems (and many other as­ from a computer-resident catalogue. Logging of the data from the other wind pects of the ND). Miguel Albrecht will soon install this and temperature sensors started al ready Because the altitude axis of the ND is catalogue on a server at La Silla so that some time ago. used in a similar fashion as the rotators,

17 the behaviourofthe altitude motors is now also being monitored. VL T Insurance Contract Signed

Automatie Display of Ineoming Images for Remote Observers

A few fellows and students in Garching now provide additional support to the operation of the ND. As a first result, we can report the installation of software by Markus Kissler which, in analogy to the practice at La Silla, also in the remote control system automatically displays newly arriving images for inspection and further analysis with MI DAS. The convenience of this new feature has been appreciated not only by the ND Team but, more importantly, also by various remote observers.

Eleetronie Operations Report System

The NDTeam has forquite some time been using acommercial system (Razor) for the tracking of technical and At a eeremony in Paris on April 11 , ESO signed an all-risk insuranee poliey for the VL T operational problems. Initially, the inten­ projeet with AGF, one ofthe leading Freneh insuranee eompanies. The poliey was also tion had been to use the same system also signedbyMessrs Fauehere & Jutheau, the insuranee brokerwhohas worked wlth ESO for the nightly operations reports by the on this eontraet. observers. However, because of the way the system had been set up and a variety of technical problems with Razor, this was never implemented. Meanwhile, the technical problems could be solved, and we thankJoseph Schwarzforhisconstant support. A second database and form for night reports have also been prepared, and the system is now in routine operation. It serves us as an eHicient and convenient source of references by means of which for instance a problem can be traced back in time and all relevant staH (plus visitors) at La Silla and in Garching can share the database.

Teehnieal Feasibility Checks of NTT Proposals

For the first time, the NTT Team has altempted to perform technical feasibility checks of all observing proposals re­ ceived for the NTT during Period 56. This was a laborious exercise. Only in a small percentage of the projects were potential problems noted; they were broughtto the attention of the Observing Programmes Commiltee for evaluation. For us, the more important aspects were (i) to obtain The signing eeremony, whieh took plaee at the AGF Headquarters, was preeeded by an early overview of the operational a weil attended press eonferenee, featuring an interesting eomposition ofjounalists requirements on us during the next period speeialising in seientifie and in finaneial malters. and (ii) the opportunity to assess under realistic conditions what kind of informa­ tion will in future be required for observa­ tions to be carried out in service mode.

The NTTTeam e-mail address: [email protected]

18 SCIENCE WITH THE VLT Topical Astrophysical Problems on Massive Stars for VLT Observations

A. MAEOER and G. MEYNET, Observatoire de Geneve, Sauverny, Switzerland

1. Introduction about 5 M at the time of supernova present 0 stars. Typically, an 07V star explosion (cf. Maeder and Conti, 1994). provides 1049 Lyman photons per second. According to current wisdom, lumi­ This evaporation is due to the stellar Refinements of this figure according to nous matter represents only about one winds in the various phases, OB stars, metallicity, upper mass limit and initial per cent of the mass in the Universe. supergiants and WR stars. Such low final mass function (IMF) slope are feasible. Among this one per cent, massive 0 and masses are weil supported by the WR The number of WR stars, when present, B stars contribute to about two thirds of luminosities and by the study of WR can also be inferred from the integrated the optical light of galaxies and thus play masses, and they have major conse­ spectrum; the broad emission line He 11 a key role in the exploration of the visible quences for all stellar properties: life­ 4686 of WR stars remains visible with an universe with the VLT. Their high lumi­ times, luminosities, evolutionary links, equivalent width of a few Adespite the nosities make them visible at large wind compositions, chemical evolution, large dilution effect by billions of other distances, either individually or by spec­ supernova progenitors and the nature of stars. In particular, the ratio of the late WN tral features in the integrated spectrum final remnants. stars (WNL) to the number of 0 stars is of galaxies. Let us note that if one measured by the ratio of He I1 4686/Hß. assumes for the VLT a limiting magni­ 2. A Remarkable Case: The Even further, the ratio of WN to WC stars tude f - 28.5 (for S/N - 2 in 10 hours, Populations of Massive Stars (i.e. of those stars exhibiting products of see Fusi Pecci et al., 1994), then 25 M, in Starbursts CNO burning to those with products of stars will be detected up to distances He burning) can be estimated from the Superior to -70 Mpc, a value wh ich Starbursts are giant events of star ratio of the He 11 4686 to CIlI/IV 4650 shows how large the volume is where formation in galaxies, they may involve lines. A test of this method has been massive stars can be observed. Massive regions with masses as large as 105 performed by Vacca (1991) using 30 Dor stars are the main sources of UV and the Orion Nebula (cf. Kennicutt, 1984; in the LMC. The number of 0 stars in a ionising radiation, also they power the Leitherer, 1991). Classic H 11 regions like fjeld of T x T centred on R136 was found far-IR luminosities of galaxies through the Orion Nebula contain only a few to be about 400 (cf. Parker and Garmany the heating of dust. Due to these ionising stars and have a total mass of 1993). The analysis of the spectrallines in properties and their short lifetimes, mas­ several tens of solar masses, while giant an integrated trailed spectrophotometric sive stars are conspicuous tracers of star H 1I regions, like 30 Dor, have about exposure of about the same area led to an formation at large distance in the Uni­ 300 to 400 ionising 0 stars and a total estimate of some 330 0 stars, a result verse. mass of about 6 105 M,.. They are even wh ich inspires a certain confidence in this Massive stars so much modify the dwarfed by starburst galaxies, like M82, procedure (cf. Vacca and Conti, 1992). physical conditions and the dynamics of Arp220, NGC6240, NGC7714wherethe Such li ne ratios, wh ich give in turn the ambient interstellar medium that they total mass involved in the burst isa few 108 ratios of star numbers, are extremely influence the process of star formation in M". The luminosity associated to star­ informative on the properties of star­ galaxies. In addition to radiation, massive bursts, like in M82, is as large as 7 10IOL" bursts, like the star formation rate (SFR), stars are also the main source of (cf. Rieke 1991). Such a value leads to the age and duration of the starbursts, the mechanical power in galaxies due to important "evolutionary corrections" on IMF and the metallicity Z. Generally, a supernova explosions and to the winds of the luminosity of distant galaxies and we higher Z leads to higher mass-Ioss rates Wolf-Rayet (WR) stars, which on the need to correctly appreciate the frequen­ for OB stars and supergiants, since there wh oie are of comparable importance. cies and intensities of the starburst are more spectral lines and thus more Massive stars with M ~ 10M" are fast phenomena, especially at large redshifts, momentum transferred by radiation. nuclear reactors, they are thus the main through the Universe both for cos­ These higher mass-Ioss rates generally Contributors to the nucleosynthesis of mological reasons and for the study of the lead to larger numbers of WR stars at heavy elements and also playa leading early evolution of galaxies. higher metallicities (cf. Maeder and role in the chemical evolution of galaxies. Fortunately, there are several ob­ Meynet, 1994). The criticalline ratios will Thanks to its high imaging and spec­ servable signatures of starbursts: the UV be observable in distant galaxies with the troscopic capacities, the VLTwill allow us light, the nebular lines, the Ha emission, VLT, they will provide very useful to observe massive stars and their effects the He 11 4686 emission line, the far-IR information on the process of star in new distant environments, like early emission, the CO line, etc. Even when a formation through the history of the evolution of galaxies, starbursts, galactic galaxy cannot be resolved into stars, it is Universe. Also, we may hope that further nuclei and star formation processes possible to learn about several properties spectral features containing information around quasars. of massive star populations (cf. Kunth and on other kinds of luminous stars like AGB A major change has occurred in our Sargent 1981; Arnault et al. 1989; Vacca stars, red supergiants, LBV, etc. will understanding of massive star evolution and Conti 1992).lndeed, the f1ux ofthe Ha become accessible with VLT observa­ in the last 10 years. The point is that these or Hß lines in the integrated spectrum of tions. All these features should be stars nearly fully evaporate during their astarburst provides an estimate of the intensively searched for and studied, and evolution. As an example, an initial 60 M,) total number NLyc of Lyman continuum the infrared spectra from the VLT might be star of solar composition will only host photons, and NLyc indicates the number of particularly useful for this purpose.

19 and the Galaxy show no systematic change of G with metalIicity Z, but the possibility that G is influenced by the INSTANTANEOUS BURST, Z=0.008 intensityofthe SFR is not unlikely (cf. also .-.. > Scalo, 1989). 0 Z These examples illustrate the high '-.. WR/O u ""/---...... ---- potentialities of future spectroscopic • -1 1 z 1 studies on massive stars with the VL1. But 1 we/o Öl _I further information could also be 1-..., \ ~ \ Z / '--\ \ obtained from other lines, forexample the \ ;, / \ Si IV and C IV UV lines in 0 stars (cf. ~ 1 '- 1 -..., \ Leitherer and Lamers, 1991; Mas-Hesse z \ \ and Kunth, 1991) or from far-IR emission 1 WNL/O \ /'" 1 -2 1 by the dust in H 11 regions. HST ob­ z :1 "" ./ / 1 servations at 2200 of the WR galaxy He l:l()= "-, : A -0 1 1 2-10 by Conti (1994) reveal knots of ....:l \: -4 1 1 \11--- ./ \ intense star formation. The application 1 WNE/O of the method mentioned above for 1 :\ I 1 estimating the mass and age of the 1 : \/ - 3 L-...L--J....-----lL--'----"-----'-....l...--4L---'---J....-'---'---5:------'--..o-J..-'----' knots has led these authors to the 3 fascinating result that one is just ob­ Age (in 106 years) serving globular clusters at birth in this Figure 1: Evolution as a function of time of the number of WR, WNL (Iate WN), WNE (early WN) galaxy, since these knots have a mass of and WC stars divided by the number of O-type stars in the case of an instantaneous burst a few 105 M" and ages certainly below occurring at time t =O. The stellar models used are those of Meynet et al. (1994) for Z =0.008. 107 years. 2 The IMF used is dN/dM 0< M- .j Aprerequisite before interpreting any observation of massive stars in distant galaxies is that the models weil apply to the Milky Way and to nearby galaxies. It 3. SFR, IMF, Age, Duration and SFR, since at this time the WR stars from is weil known that large differences in Metallicity in Starbursts the burst have not yet appeared. From the massive star populations exist wh ich can information provided by the WNL/O and be interpreted by means of stellar models One big hope for future observations WN/WC ratios we can make an estimate (cf. Maeder and Meynet, 1994). For of massive stars with the VLT is to of the age of the burst and put some example, the number ratios WR/O and disentangle the various parameters bounds on its duration in a given area. WC/WN differ by factors 10 and 20 characterising the starbursts. As a first These number ratios are also sensitive to respectively between the SMC and inner example let us examine the didactical, the slope ofthe IMFfora given metaliicity regions of the Milky Way or of M31. but not unrealistic case of an instantane­ Z. Comparisons have shown that these ous starburst. According to recent mod­ An interesting application, which large differences could be accounted for els (cf. Maeder and Conti, 1994; Meynet, should be extended in future to more by metallicity effects wh ich enhance the 1995) we may distinguish four different distant galaxies, is that of H 11 or WR mass-Ioss rates and in turn influence epochs in the initial evolution of a star­ galaxies (cf. Kunth and Sargent, 1983; stellar evolution. burst (see also Fig. 1) Conti, 1991 a); these are emission-Iine 1. 0 phase: for an age t :::; 2 Myr, only galaxies which show a broad emission o stars are present, giving rise to H I1 feature at He 114686. This line is also often 4. SN-Dominated Galaxies and regions without WR features. shown by AGN's, but in WR galaxies the Nucleosynthesis 2. WNL phase: from t =2 to 3 Myr, a nebular lines are narrow, as caused by large number of WNL stars are present. stellar ionising radiation in giant H II An interesting result of starburst 3. WN and WC phase: from t =3 to regions. Some of these WR galaxies models to be explored is that there should 7 Myr, the various subtypes of WR stars show signs of collisions, mergers, jets, be a phase dominated by supernova (SN) (WNL, WNE, WC) coexist with fractions some are double, one is an IRAS galaxy. explosions in the subsequent life of a depending on mass loss and metaliicity Z. These H II orWR galaxies generallyshow, starburst. A rule of thumb (cf. Meynet, 4. LateOphase:fort'?7 Myr, WR stars as derived from the line ratios, large 1995) is that one should have about one have disappeared, but up to 10 Myr there excesses of WR stars with respect to 0 SN per century for the birth of 13,500 0­ are still 0 stars. They produce H I1 regions stars; Figure 2 shows the model type stars. Applying this to the powerful with nebular emission lines, but the predictions by Maederand Meynet (1994) starbust galaxy NGC 1614 in the list of equivalentwidth oftheHßli neforexample and Meynet (1995) compared to the Vacca and Conti (1992) one could expect should be much smaller than for the early observations by Vacca and Conti (1992); about 12 SN per century. In the extreme o phase. see also Maeder and Conti (1994). Burst case of the IRAS galaxy 01003-2238 It has to be noted that the ratio WNUO models with two different IMF slopes are (Armus et al., 1989) one would have is much larger in astarburst (up to 6 times shown, where the IMF is defined by dN/ about 45 SN per century. Do we observe for solar Z) than in the case of constant dM =AM (-1). Figure 2 shows that the such "SN galaxies" in a consistent way SFR (see also Fig. 2), where there is observed WNL/O ratios in these starburst with existing starbursts? If not, why? If always an equilibrium between stars galaxies are above the predictions for yes, the effects of such extreme situa­ entring and leaving the WR stage. constant SFR and in nice agreement tions on the interstellar medium and on Surprisingly also, if an observed H II with burst models. The uncertainties are the chemical enrichment of galaxies are region consists of a burst plus a region of for now probably too large to enable us to fascinating topics about which almost lower but constant SFR, we shall observe infer r values for starburst regions. In this nothing is known at the present time. overthe 2 Myrafterthe bursta much lower context it might be worth recalling that The chemical evolution due to star­ WR/O ratio than in the case of constant comparisons of clusters in the SMC, LMC bursts is also an unexplored subject.

20 High-resolution spectroscopy from the VLT could bring a lot of new information. The key effect in the interpretation of past chemical abundances rests on the fact o HIGH MASS LOSS RATE that massive stars have short lifetimes. In MODELS • the early phases of galactic evolution, • only massive stars contribute to the chemical enrichment due to their short lifetimes. As time goes by, smaller stellar masses come into the game: firstly, only SNII contribute to the chemical enrich­ -1 ment (mainly in 0, Ne-Ca elements and r=l r-process elements), then appears the production of the intermediate-mass • stars (with mainly C, N and s-process elements); they are fOllowed by the r=2 contributions from supernovae SNla • (mainly Fe injection). Thus the ages at which stars of a given mass release their -2 nucleosynthetic production are generally considered as the major effect regarding the changes of stellar yields as a function oftime (cf. Matteucci, 1991). One should, ...... Const. SFR however, also take into account the fact that the chemical yields are changing with initial Z and may this way influence the -3 Inst. Burst picture of chemical evolution of galaxies (Maeder, 1992).Aschematicsketch ofthe various interplays intervening in galactic evolution is given in Figure 3. -1.5 -1 -0.5 o 0.5 In the case of astarburst the chemical [O/H] contribution of the burst is only that of massive stars above some limit which Rgure 2: Relative number of WNL to O-type stars as a function of the relative abundance ofoxygen depends on the age of the starburst. expressed in terms of[O/Hj. Filled circles are observed values as given by Vacca and Conti (1992). The open triangle represents 30 Doradus (from Vacca 1991) andopen boxes are observations from Vacca These yields are thus different from the as reported by Conti (1994). The dotted fine presents theoretical values obtained in thecase ofaconstant average yields, especiallY if starbursts star formation rate with an IMF, dN/dM

21 PROPERTIES mechanism praducing this kind of be­ OFSTAR FORMATION haviour. 7. Chemical Abundances at the Surface of Massive Stars

Very luminous and massive stars (which are of particular importance for extragalactic studies) are often too faint for detailed studies with existing tele­ scopes (see Appenzeller, 1986). The great aperture of the VLTwili enable us to obtain the high-resolution spectrograms 4 5 (R = ')J/>,'/.. between 10 and 10 ) with high 2 3 signal-to-noise ratios (S/N - 10 to 10 ) necessary to the measurement of reliable abundances at the surface of extremely bright stars in extragalactic systems of different chemical composition. These observations can then be used as standards for lower resolution studies of more distant stellar systems. A lot of very interesting observational facts concerning the surface composition of massive stars have been collected in the past years. In particular, the recent finding of Herrero et al. (1992) that all fast rotators among 0 stars show surface He­ enrichments is a very important result. It shows that rotationally induced mixing is strang enough to transport the newly­ I PHYSICS OP STAR PORMATION processed elements to the surface during a fraction of the main-sequence lifetime. STAR EVOLUTION 11 This implies that all stellar properties 111 CHEMICAL TIELDB AND BVOLUTION are influenced by this effect: lifetimes, luminosities, nucleosynthesis, etc. IV DBPOSITION OP NOWBNTUM, BNBRGY Walborn (1976, 1988) showed that EPPBCTS V OP xi ON IMP, SPR ordinary OB supergiants have He- and N­ VI BPPBCTS OP xi ON STAR BVOLUTION enrichments as a result of CNO processing. Only the small group of VII BXTBRNAL BPPBCTS ON GAS DYNAKICS peculiar OBC-stars have the normal Figure 3: Schematic illustration of the interplay between star formation, stellar population, cosmic abundance ratios (cf. also chemical and dynamical evolution of the interstellar medium. Howarth and Prinja, 1989; Herrero et al., 1992; Gies and Lambert, 1992). There are two alternatives to explain the He- and N enrichments in supergiants and both show a significant stellar population just The number ratio of blue to red bring about new problems (Maeder, on the red side of the main sequence (log supergiants is an increasing function of 1995). Teff = 3.9 to 4.1) for initial mass stars the metalIicity Z in galaxies, typically (1) The first alternative is that blue between 10 and 20 Me. The presence of increasing by a factor of 10 from the SMC supergiants are on the blue loops after a numerous stars in this part of the HR to the solar neighbourhood (cf. Hum­ first red supergiant stage where dredge diagram is quite puzzling since, according phreys and McElroy, 1984). Langer and up has occurred producing the observed to standard numerical simulations, evolu­ Maeder (1995) showed that if, at a given surface enrichments. The problems re­ tion is quite rapid at this stage (blue metallicity, it is possible to account for the lated to this hypothesis are the following: Hertzsprung gap). Several different observed blue to red ratio, by changing (a) At Z = 0.02, current grids of models possible explanations have been pro­ the mass-Ioss rates or the mixing (Arnett, 1991; Schalleretal., 1992; Alongi posed in terms of enhanced opacities mechanism, no assumption on stellar etal., 1993; Brocatoand Castellani, 1993) (Nasi and Forieri, 1990), of scatter in model physics explored so far is able to only predict blue loops for masses equal mass-Ioss rates (Meynet et al., 1994), cope with the general trend. All models to or lower than 15 M.. At higher masses extended atmospheres and binaries do the opposite, i.e. many more red there are no blue loops and no predicted (Tuchmann and Wheeler, 1989, 1990). supergiants are predicted at high Z than enrichments. (b) Even on the blue loops Some incompleteness effects and prob­ at low Z. Langer and Maeder (1995) the predicted enrichments do not seem lems of calibration relations between B-V suggest some connection of the blue to high enough to account for the obser­ and log Teff mayaiso be responsible for at red problem with internal mixing. Indeed vations. least part of the problem. However, our they found that a restricted scheme for (2) Another possibility is that the feeling is that no good solution has been mixing (Ledoux) looks generally better at mixing responsible for the surface found yet and that new evolutionary low Z, while an extended scheme might abundances of these supergiants oc­ tracks are needed to account for this be better at high Z. The question is of curred al ready on the main sequence. important fact. course to know what is the physical We may, for instance, assume that the

22 abundances of 22Ne at their surface as a O,B F,G SUPG WNL WNE wc resuft of the transformation, at the IH 4He beginning ofthe He-burning phase, ofthe ° 4He 14N lett by the CNO cycle (Maeder, 1990). However, the infrared observations by 160'" INITIAL 60Ms ' Z=O.02 Barlowetal. (1988), forthe WC8stary-Vel 22Ne give Ne/He =0.005 (in mass fraction), i.e. 160 aN r·-·-·-----·------·----- .----' at least 6 times less than predicted. -2 (However, this is twice the solar ratio). Let - -- - r2C~ - -- '""'~ 26Mg us note that the issue of this problem does ~,... ~ 16 ...... -- not only concern a detail of stellar 0 ~26Mg 0-1 I -- - -23N- - - - - structure but is also important for the ----~---- nucleosynthesis of s process elements in massive stars, since the 22Ne is the major 21Ne neutron source in these stars. It has also ----- some implications on the understanding of the high values for the isotopic ratio 22Ne/2°Nefound in cosmic rays, which has -6 been interpreted as the signature ofa WR contribution (Maeder and Meynet, 1993). At present, observations of mass loss, w ~ ~ W of ejected shells, of surface chemistry, of Mr/MS LBV and WR properties are rather scarce Figure 4: Changes of surface abundances (in mass fraction) versus the remaining stellar mass and much safer conclusions on the last in solar units for a 60 M(') model with Z =0.02. At the top, the corresponding evolutionary stages stages of massive star evolution in are indicated. galaxies of different metallicities are still very much needed. effects producing enriched 0 stars (see phase. More observations are required 9. The Supernova 1987A, above) are also active in lighter lower to go ahead with this hypothesis. Stellar Remnants mass stars and that they lead to surface Among the stars with extraordinary enrichments in supergiants resulting mass outflows, WR stars certainly occupy According to McCray (1993), the light from a larger range of initial masses. a first-rank position. These stars are curve of SN 1987A, powered by 44Ti Whatever the solution, there is no nowadays considered as bare cores lett decay (t44 = 78 yr), should remain steady doubt that, as for the problem of the blue over from massive stars by mass loss; around m :::: 19 for decades and thus be to red supergiant ratio, mixing processes however, the exact cause of this extreme quite weil observable with the VLT. There playa key role in understanding these mass loss is still unknown. Recent works is no doubt that this zone of the sky will observations. have succeeded in explaining the large reserve further surprises when studied observed changes of the number ratios with this powerful instrument. First of all WR/O, WC/WN in galaxies of different one could be able to test the hypothesis of 8. luminous Blue Variables and metallicities (cf. Maeder and Meynet, the presence of a companion. Indeed, Wolf-Rayet Stars 1994). These changes are due to the some authors (Soker and Livio, 1990; larger mass-Ioss rates at higher Z. Also, Podsiadlowski, 1992) have invoked the Some of the most luminous stars the models were able to explain the possibility that Sk-69 202 was a binary. If have sporadic, violent mass-Ioss events observed trend of WC9, WC8 and WC7 to not disrupted by the explosion, its mass whose causes are not understood. These be located in inner regions of the Galaxy. must be inferior to 2.5 M.., otherwise we evolved hot stars are called luminous blue The sequence of WR stars nicely should see its light in the optical continu­ variables (LBV) and their instability may corresponds to a sequence of more um. Anothervery interesting question will shape the appearance of the upper HR advanced chemical processing becom­ be to know what kind of remnant this diagram (Humphreys and Davidson, ing visible at the stellar surfaces (see Fig. explosion has lett, a neutron star or a 1995; see also Maeder and Conti, 1994). 4). The standard models (cf. Schaller et black hole? An optical pulsar might be Presently, these stars are interpreted as al., 1992) predict an abrupt transition from detected (with mv ::; 23) if it has a ~ 34 S 1 a short stage in the evolution of massive WN to WC stars, since the He core is luminosity LoPI 3 x 10 ergs - (McCray, stars with initial masses M > 40 MD . A growing and building up a steep chemical 1993). It is also expected that when the likely scenario is (cf. Maeder and Conti, discontinuity at its outer edge. Thus the supernova ejecta will interact with matter 1994). standard models predict almost no « 1%) expelled by the progenitor, very spectac­ stars with intermediate characteristics of ular outputs will be produced. Luo et al. star ~ Of/WN ~ LBV ~ Of/WN o WN and WC stars. However, there are (1993) estimate that a shock will strike ~ LBV ... ~ WN ~ WC about 4% of such intermediate stars the ring of matter which surrounds the Are these violent eruption stages (Conti and Massey, 1989). It has been supernova at a radial distance of - 0.7It­ responsible for some of the rings shown by Langer (1991) thata mild mixing yr, in 2004 ± 3. At that time, strong radio, observed around WR stars? Marston et like that produced by semiconvective optical and UV emissions are predicted al. (1994) have found that of the 145 diffusion can account for these stars. The to occur (Luo and McCray, 1991; Suzuki galactic WR now surveyed, more than a existence of WN/WC stars undoubtedly et al. 1993; Luo et al., 1992). After the quarter have associated ring nebulae. shows that an extra-mixing process is at initial impact, the ring will become a bright Several WR are shown to be surrounded work, even during the short He-burning source of sott X-rays and infrared con­ by multiple rings (2 or 3) wh ich suggests phases. tinuum. the occurrence of multi-ejection stages Exceptatvery low metallicity, WC stars The importance of the mass limits for as is thought to occur during the LBV are expected to present very high level neutron stars and black holes for the

23 chemical evolution of Galaxies has been than in the visible, Cepheids, in this Langer N., Maeder A. 1995, A&A, in press. recently discussed by Maeder (1992). wavelength range, exhibit a tighter Leitherer C. 1991, in Massive Stars in The yields in heavy elements strongly period-Iuminosity relationship which is Starbursts, Ed. C. Leitherer et al., Cam­ bridge Univ. Press p. 1. decrease with the lowering of MSH ' the less metallicity-dependent (Moorwood, LeithererC., Lamers HJGLM. 1991, ApJ 373, 89. lowest mass limit for black hole formation, 1986). Luo 0., McCray R. 1991, ApJ379, 659. since then more heavy elements are Luo 0., McCray R., Slavin J. 1992, Bull. Am. swallowed by black holes. Comparisons Astron. Soe. 24, No. 4, p. 1244. with observational data, in particular with 11. Conclusion Maeder A, 1985, A&A 147,300. Maeder A. 1990, A&AS 84, 139. the DYI DZ ratio suggest a value of M SH A universe without massive stars Maeder A. 1992, A&A 264, 105. around 20 or 25 M. .. The dependence of Maeder A., Conti PS. 1994, Ann. Rev. Astron. would be nearly adead universe with such the yields and supernovae types on Astrophys. 32, 227 long evolutionary time scales that it would Maeder A, Meynet G. 1994, A&A 287,803. metallicity and on the value of MSH is a key problem in stellar astrophysics with many appear quite unevolving. Indeed, even if Maeder A. 1995, in Astrophysieal Applieations of massive stars are quite rare objects, they Stellar Pulsation, lAU Coll. 155, in press. implications. Much theoretical and ob­ MaederA., Conti P. 1994, Ann. Rev. A&A 32,227. servational work is still needed in this are nevertheless the main engine of many active processes, as stated in the Maeder A., Meynet G. 1993, A&A 278,406. area. Maeder A, Meynet G. 1994, A&A 287,803. introduction. Their role in the study of star MarstonAP, Yocum D.R., Garcia-SeguraG., Chu formation and evolution of galaxies, as Y.-H., 1994. ApJS95, 151. 10. Massive Stars in Galaxies as weil as the many problems wh ich still Mas Hesse J.M. 1992, A&A 253, 49. Distance Indicators need to be clarified on their internal and Mas Hesse J.M., Kunth D. 1991, A&A Suppl. 88, external physics, make massive stars 399. As al ready emphasized above, mas­ Massey P, Lang C.C., Degioia-Eastwood K., first-rank goals for VLT observations. sive stars are fantastic tools to explore the Garmany C.D. 1995, ApJ438, 188. remote universe. The VLT will be able to McCray R. 1993, ARAA 31, 175. Matteucci F 1991, in "Chemistry in Space", Erice observe long-period Cepheids up to a References School, Ed. M. Greenberg, KluwerAcad. Publ. distance of 70 Mpc and red supergiants p.1. up to 700 Mpc. Type la supernovae will Alongi M .. Bertelli G., Bressan A., Chiosi C., Meynet G. 1995, A&A in press. be detected up to 7000 Mpc (Moorwood, Fagotto F et al. 1993, A&AS 97,851. Meynet G., Maeder A, Schaller G., Schaerer 0., 1986). Appenzeller 1.1987, in ESQ's Very Large Charbonnel C. 1994, A&A 281, 638. Cepheids, because of their impor­ Teleseope, ESQ Con!. and Workshop Moorwood AF.M. 1987, in ESQ's Very Large Proceedings 24, S. D'Odorico and J.-P. Teleseope, ESO Cont. and Workshop Pro­ tance in the determination of the Swings (eds.), ESO, p. 55. ceedings, 24, S. D'Odorico and J.-P Swings cosmological distance scale, deserve Armus L., Heckman T. Miley G. 1988, ApJ326, (eds.) ESO, p. 55. particular attention (cf. the recent de­ L45. Nasi E., Forieri C. 1990, Astrophys. Sp. Sei. 166, tections of Cepheids in galaxies of the Arnault P., Kunth 0., Schild H. 1989, A&A 224, 229. Virgo cluster by Pierce et al., 1994, 73. Palla F, Stahler S.W, 1992, ApJ 392,667. Arnett D. 1991, ApJ 383,295. Pierce M.J. et al. 1994, Nature 371, 385. Freedman et al., 1994, Saha etal., 1995). Barlow M.J., Roche P.F., Aitken D.K. 1988, Podsiadlowski P 1992, PASP 104, 717. Since different modes of pulsation follow MNRAS 232, 821. Rieke G. 1991, in Massive Stars in Starbursts, Ed. different period-Iuminosity relations, it is Blaha C., Humphreys R.M. 1989, AJ98, 1598. C. Leitherer et al., Cambridge Univ. Press, p. crucial to obtain accurate distances to Brocato E., Castellani V. 1993, ApJ410, 99. 205. be able to discriminate between these Churchwell E. 1991, in The physies of Star Saha A., Sandage A, Labhardt L., Schwengeler Formation and Early Stellar Evolution, H., Tammann GA, Panagia N., Macchetto F.D. modes. The effects of the chemical NATO ASI Series, 342, C.J. Lada & N.D. 1995, ApJ438, 8 composition has also to be fully un­ Kylalis (eds.), KluwerAcademic Publishers, ScaloJ., in Windowson Galaxies, Ed. G. Fabbiano derstood. In this context it is interesting to p.221. et al., Kluwer Acad. Publ. p. 125. emphasize that the Cepheid models Conti P.S.1991, ApJ377, 115. SchaUer G., Schaerer 0., Meynet G., Maeder A. predict different surface He-contents and Conti P.S. 1994, Spaee Sei. Rev. 66, 37. 1992, A&AS96, 269. N/C ratios according to metallicity Z and Conti P., Massey P. 1989, ApJ 337,251. Schneider D.P., Schmidt M., Gunn J.E. 1991, Fitzpatrick E.L., Garmany CD. 1990, ApJ 363, Astron. J. 102, 837. masses (cf. Schaller et al. 1992). The 119. Soker N., Livio M. 1989, ApJ339, 268. difference in helium content, Ys =0.243 at Freedman W.L. et al. 1994, Nature 371, 757. Stahler S. 1983, ApJ274, 822. Z =0.001 compared to Ys =0.33 at Z = Fusi Pecci F., Cacciari C., Ferraro FR., Suzuki T., Shigeyama T., Nomoto K. 1993, A&A 0.020, should have some consequences Gnatton R., Griglia L. 1994, The Messenger 274,883. for the driving mechanism and thus 77,14. Tuchman J., WheelerJ.C. 1989, ApJ344, 835. Garmany CD., Conti P.S., Chiosi C., 1982, Tuchman J., WheelerJ.C. 1990, ApJ363, 255. for the amplitudes of Cepheid pulsations ApJ263,777. VaccaWD.1991, WRstarsinMilkyWay, theLMe at different Z. Such differences of am­ Gies D.R., Lambert D.L. 1992, ApJ 387,673. andEmission-line Galaxies, Ph.D. thesis, Univ. plitude distributions between Galaxy, Hamann F., Ferland G. 1992, ApJ391 , L5. 01 Colorado, Boulde. LMC and SMC are known to exist (cf. HerreroA., Kudritzki R.P., Vilchez J.M., Kunze Vacca WO., Conti PS. 1992, ApJ401,543. van Genderen, 1978) and we may sus­ 0., Butler K., Haser S. 1992, A&A 261,209. van Genderen AM. 1978, A&A 65, 147. Howarth 1.0., Prinja R.K. 1989, APSJS 69, pect a relation with the different helium Walborn N.R. 1976, ApJ205, 419. 527. Walborn N.R. 1988, in AtmosphericDiagnosties of contents. Humphreys R.M., Davidson K. 1994, PASP StellarEvolution, lAU Coll. 108, Ed. K. Nomoto, The possibility for the VLT to perform 106,1025. Springer Verlag, Berlin, p. 70. infrared observations will also be Humphreys R.M., McElroy D.B. 1984, ApJ 284,565. important in this context. Indeed, in --- -_._------addition to the factthat, in the infrared, the Kennicutt R.C. 1984, ApJ287, 116. Kunth 0., Sargent W.L.W 1981, A&A 101, L5. standard distance indicators are subject Kunth 0., SargentW.L.W1983, ApJ273, 81. G. Meynet to much lower extinction uncertainties Langer N. 1991, A&A 248, 531. e-mail: [email protected]

24 , REPORTS FROM OBSERVERS

Monitoring of Active Galactic Nuclei: the Why and the How

D. ALLOIN1, M. SANTOS-LLE01, G. STlRPE2, and B.M. PETERSON3

1Departement d'Astrophysique et de Cosmologie, URA 173 CNRS, Observatoire de Paris-Meudon, France; 20sservatorio Astronomico di Bologna, Italy; 3Department ofAstronomy, The Ohio State University, Columbus, Ohio, USA

1. Introduction the inner structure of AGNs is provided by collaboration (Ulrich et al., 1984; Clavel et measuring in detail how the emission-line al., 1990). One of the major surprises of Over the past twenty years, ground­ fluxes change in response to changes in the monitoring campaigns of the eighties based optical observations as weil as the continuum flux. The broad emission was that the BLR seemed to be an order ultraviolet, X-ray, and y-ray observations lines respond with small but measurable of magnitude smaller than the value Irom space-borne telescopes have re­ time delays (days to weeks) to variations generally predicted by photoionisation vealed the variable nature 01 the conti­ of the central continuum source, making equilibrium calculations. This conclusion nuum and emission lines in the spectra it possible to use the technique of demanded even denser sampling for 01 active galactic nuclei (hereafter re­ "reverberation mapping" to probe the AGN variability programmes. lerred to as AGN). Variability is now structure and kinematics of the BLR. In In 1987, two successive workshops in recognized as one of the distinctive general, the narrow lines do not vary in Segovia and Atlanta featured lively dis­ leatures of these important but poorly flux since the size of the NLR is usually cussion of results obtained from AGN understood objects. too large to provide a coherent response variability studies. It became apparent to to changes in the level of the continuum the community that the goals of spectro­ flux. scopic monitoring programmes could in 2. The Why of Monitoring The fundamentals of reverberation fact be achieved only if sufficient observ­ Campaigns mapping were described by Blandlord ing time could be devoted to such an and McKee (1982), but it has been only approach. 2. 1 Some basics over the last five years or so that the Cooperation of observers on a scale first tentative applications of this tech­ that was unprecedented in extragalactic In the study 01 AGN, variability affords nique to real AGN have been possible, astronomy, i.e. with very large collabora­ a potentially valuable probe of the as severe conditions on the amount and tions involving around 100 astronomers, properties of both the continuum source quality of the data have to be met (Pe­ became a necessity. itself and the broad-line emitting region terson, 1994). To deal with a collaboration of this surrounding it. Some experiments, undertaken by the size and a highly time-constrained pro­ Knowledge of the continuum variabili­ "International AGN Watch" collaboration, gramme, new ways of working and ty pattern in different wavebands from y­ have been conducted in part with tele­ cooperating had to be invented. The ray to radio wavelengths can provide a scopes at the European Southern Obser­ International AGN Watch was therefore way to probe the various physical pro­ vatory (ESO), and these form the subject established with the goal of focusing cesses at the origin of the continuum of this report. attention on a few AGN for intensive emission. The variation time scales in Similar programmes, albeit with a monitoring efforts and maintaining com­ particular give some indication of the size sometimes different overall emphasis, munication among the various individu­ of the emitting regions, an indirect clue to have been undertaken by other informal als and groups that carried out the actual the likely emission mechanisms. Of organisations du ring the same time observations. A key factor in the success potentially greater interest is the possibil­ Irame. For example, the European con­ of these efforts has been the ability to ity of eventually measuring time lags sortium LAG ("Lovers of Active Galax­ communicate and exchange information between continuum variations in different ies") wh ich was initiated by the late MV promptly via modern computer networks. wavebands because this can tell us about Penston, has carried out a spectroscopic The role of the AGN Watch has been the connection between various mecha­ and photometric monitoring of several multifold: (a) to define the scientific nisms producing continuum photons in AGN on the Canary Islands telescopes questions to be addressed, conceive the these systems. within the framework of the CCI 5% observational projects and coordinate In the framework of the so-calied international time programme (Robin­ the submission of the appropriate observ­ standard model, based on a massive son, 1994). ing proposals, (b) to ensure that data are black hole and accretion-disk system, we collected in a manner consistent with the assume that accreting material is distrib­ 2.2 A bit of recent history scientific goals, (c) to reduce the observa­ uted throughout the line-emitting regions: tional data and make these data sets the broad-line region (hereafter BLR) and In the early eighties, a number 01 available to the entire community, and (d) the narrow-line region (NLR), wh ich are groups involved in AGN studies under­ to perform the measurement and analy­ Somewhat arbitrarily distinguished by the took ultraviolet and optical monitoring sis of the data and publish the primary width of the lines they emit (ranging from programmes in an effort to probe the scientific results. as much as a few 10,000 km/s for the physics of AGN (for a review, see It was decided that detailed and broadest lines to only a few 100 km/s for Peterson, 1988). We note in particular the model-dependent interpretation would the narrow lines). An important probe 01 results of the so-calied NGC 4151 be left to individuals or sub-groups of the

25 16 ultraviolet spectra at a higher rate than in the original programme, once every two days with IUE between March 16 and ,...... May 27, 1993; during the second half of I 1I 0« 1 this campaign, HST spectra were ob­ N 14 I tained with an even higher frequency, 8 once per day. Detailed results of the ü h campaigns on NGC 5548 can be found I in Clavel et al. (1991), Peterson et al. Vl IIIHII It (1991), Korista et al. (1995) and refer­ QJ) 12 l-o r ences therein. The main conclusions llJ I I reached are as folIows: ~ I 1. The ultraviolet and optical continua ...... 0 vary with little, if any, phase difference '--' 1 between them. The continuum becomes u10 r.r... ~ 1 bluer as it becomes brighter and the shorter-wavelength continuum bands show sharper variations. 2. The variations of the highest ionisa­ tion lines (He 11, NV) lag behind the 1300 variations of the ultravioiet continuum by ...... N slightly less than 2 days, implying an I 8 inner radius of somewhat less than 2 Ü1200 light-days for the BLR. Its outer radius, ~ ~ 11 from the CIII] and Balmer lines, is I l Vl li 1 1 1 somewhat larger than 20 Iight-days. QJ) q\#lt I I 3. There are some indications that the l-o 11 llJ1100 ~ I jl 11 higher radial-velocity gas (line wings) ~ I I fIlII responds more rapidly than the lower I radial-velocity gas (Iine core), suggesting 0..... 11 "-" 11 1 a virialised BLR c10ud system. ..--..1000 co. I ::c: ~l 3.2 Where ESO comes on the "-" I r.r... tl I stage, the NGC 3783 campaign 900 In order to improve our understanding of the size and structure of the BLR and to test the generality of the NGC 5548 results, it was deemed to be desirable to Figure 1: The light-curves ot the optical continuum (top panel) and ot Hß (10 wer panel) trom the carry out similar programmes on other AGN in NGC 3783 during the ESO campaign. AGN in order to map the AGN luminosity vs. BLR size plane. Therefore, two other targets with different absolute luminosities were se­ AGN Watch collaboration, as weil as to ground-based and space-based obser­ lected, NGC 3783 (AGN Watch 11) and other interested parties. The AGN Watch vatories. Fairall 9 (AGN Watch IV), both observ­ data are at the disposal of the entire able from the southern hemisphere. community once the primary scientific 3. 1 First experience, the NGC 5548 These AGN Watch campaigns relied results have been published by the campaign heavily on ESO telescopes for the collaboration. ground-based component. The first AGN Watch project was an The AGN Watch campaign 11 was set eight-month monitoring campaign (AGN up to monitor NGC 3783 with IUE for 69 The How Watch campaign I) on the Seyfert 1 epochs from December 21, 1991 to July gaiaxy NGC 5548. IUE observations 29, 1992, once every 4 days for the first As many of the strongest and most were made once every four days be­ 172 days and once every 2 days for the important broad lines in AGN spectra are tween December 14, 1988 and August 7, final 50 days. Simultaneous optical and located in the ultraviolet domain, space­ 1989, for a total of 60 epochs. near-infrared observations were collect­ based observations are critical for under­ Ground-based observations were col­ ed from ESO and CTIO (Chile), CASLEO standing the BLR. Therefore, the initial lected with various telescopes in the (Argentina), Lowell Observatory (USA), focus of the AGN Watch efforts was UV northern hemisphere during this entire Vainu Bappu Observatory (India) and spectroscopy using the International UI­ period, and the ground-based compo­ SAAO (South Africa). The ground-based traviolet Explorer (IUE) and, later, the nent of this programme is in fact still campaign started on December 3, 1991 Hubble Space Telescope (HST). Howev­ continuing. The success of this original and was completed on August 9, 1992. er, concurrent observations of the optical programme (see Peterson, 1993, for a An application for the ESO observing lines and continuum are essential for a detailed summary) led to a similar programme was submitted through the complete description of AGN behaviour, programme of NGC 3783 (AGN Watch 11, normal Observing Programmes Commit­ and therefore each of the AGN Watch described below), and to a follow-up tee (OPC) channel, with special require­ campaigns was organised to ensure that programme on NGC 5548 (AGN Watch ments regarding the dates to be sched­ adequate data would be obtained at both 111) using both IUE and HST to obtain uled and the time-slot to be allocated to

26 the programme, i.e., once every 4 nights Year, it was possible to arrange a nearly observation times, with special care being (concurrent with the IUE observations), 2 simultaneous multi-wavefength snapshot taken to maintain the regular sampling hours of time placed such that NGC 3783 of NGC 3783which included observations that is desirable in these monitoring would be at its meridian transit (in orderto from GRO, ROSAT, Voyager, IUE, HST, programmes. The AGN Watch observa­ minimise the air mass). optical and infrared ground-based tele­ tions were scheduled even when the The project was recommended by the scopes, and the VLA. These data were telescope was otherwise idle, and we OPC and carefully scheduled by ESO on also essential to a better understanding of greatly benefitted from this high level of the 1.5-m telescope. All Pis of other the continuum source in NGC 3783 and cooperation by ESO. regular successful 1.5-m telescope pro­ complemented the AGN Watch data sel. The on-site organisation of the cam­ posals were informed in advance by ESO The results of AGN Watch campaign 11 paign was roughly similar to that set up for about these two-hour blocks of time and are described in detail by Reichert et al. AGN Watch 11, with Dr. C. Mendes de were required to schedule their own (1994), Stirpe et al. (1994), and AIIoin et Oliveira and Dr. E.Chatzichristou being observations around this interruption. A al. (1995). We note here the salient successively responsible for the interac­ detailed handbook had been prepared for conclusions: tion with Pis of the regular proposals the AGN Watch observations, describing 1. As in the case of NGC 5548, scheduled on the 1.5-m telescope at briefly the purpose of the observations, significant variations were detected, both every AGN Watch epoch, and forthe data the experimental conditions to be strictly in the continuum and in the emission-line collection. Again, the campaign went on followed, and providing information for all f1uxes. We observe in NGC 3783, how­ smoothly as the ESO staff and most of the the necessary contact persons. A group ever, rapid f1uctuations of relatively high­ regular observers were extremely coop­ of ESO postdocs, students, and cooper­ er amplitude than in NGC 5548, while the erative and helpful. ants present at La Silla over this period of longer-term modulations appear to be Data from the ground-based cam­ time was organised under the responsi­ comparatively less weil defined. paign, from all participating observato­ bility of Dr. B. Altieri to actually take care 2. The continuum fluxes appearto vary ries, are now being reduced and will be of the AGN Watch observations. Dr. B. simultaneously in all four measured compared soon to the ultraviolet continu­ Altieri was also in charge of collecting and ultravioletloptical bands. The slope of the um and emission-line Iight-curves. Pre­ reducing all the AGN Watch data on-line ultraviolet continuum is found to vary in liminary results from the IUE campaign in order to ensure that the programme the sense that the fractional amplitude of show that Fairall 9 did vary significantly was being carried out as designed and the continuum variations decreases with and that high trequency variations are was producing useful data. increasing wavelength. superimposed on the longer-term trend. This organisation turned out to be 3. Cross-correlation analysis indicates extremely satisfactory and efficienl. The strikingly short time delays for most of the ESO staff was found to be very coopera­ strong emission lines. The peaks of the 4. Conclusions tive, wh ich was certainly one of the cross-correlation functions occur at lags primary reasons for the success of the of 0 ±3 days for He 11+0111], 4 ± 2 days for Although the observational effort de­ campaign. The observers of regular Lya and CIV, 8 ±3 days for Hß, and 8-30 manded in such campaigns is quite large, proposals had to deal with some interrup­ days tor Si 111] and CIII]. the AGN community is convinced that the tion of their own programmes and consult 4. The continuous emission of the scientific returns are sufficiently impor­ with the ESO AGN Watch team with genuineAGN in NGC 3783 appears to be tant that such campaigns are worth the regard to details of the programme in rather flat from soft y-ray to infrared trouble. Through these large-scale coor­ order to carry out both the PI and AGN wavelengths with index a '= 1. The dinated efforts, we have been able to Watch programmes as efficiently as ultraviolet and near-infrared excesses acquire data sets of high quality and Possible. We are pleased to report that we can be understood in terms of thermal reasonable homogeneity which are suit­ found a highly cooperative spirit among emission from an accretion-disk surface able for further statistical analysis. The the regular observers. Altogether, the and a hot dust component, respectively. AGN Watch data sets are available to the experience has been quite positive in our entire community, and there is no doubt relations with the ESO staff and the 3.3 A high luminosity object, that they will be used by many more European astronomical community. We the Fairall 9 campaign astronomers in the future. owe them many thanks and certainly The AGN Watch campaigns have some part of the credit for the success of AGN Watch campaign IV was devoted demonstrated that there is no delay, to the campaign. to an AGN of much larger absolute the accuracy measurable so far, between After reduction, the ESO data were luminosity, Fairall 9, which was already the ultraviolet and optical continua. This then merged with similar data sets known to exhibit long-term large-ampli­ result rather argues in favour of repro­ collected at other ground-based facilities tude variability (Clavel et al., 1989). In the cessing models and seems to rule out (Fig. 1). There was in particular a very ultraviolet, this object was observed with simple, geometrically thin, optically thick close collaboration with CTIO, as re­ IUE once every 4 days from April 30 to accretion-disk models for AGN. In at least searchers in charge of the AGN Watch at December 26, 1994. Because Fairall 9 is three AGN, it has been confirmed that the both sites were in continuing contacl. a southern source, again the 1.5-m ESO size of the BLR is an order of magnitude The NGC 3783 campaign was also telescope played a key role in the ground­ smaller than predicted on the basis of distinguished from the AGN Watch cam­ based effort. Again the standard proce­ the standard photoionisation equilibrium paign I by two important features: dure of time application through the OPC models of AGN. In NGC 3783, it is found 1. The availability of HST allowed us to channel was followed, without calling for that the BLR extends from about 1 to 2 obtain a high-resolution, high signal-to­ a key projecl. Time was granted to the light-days upwards to around 30 light­ noise ultraviolet spectrum that proved to campaign, 2.5 hours once every 4 days days, and is radially structured with the be crucial in disentangling features in the from May 2 to September 27,1994 but highly ionised material closest to the IUE data by using the HST spectrum as a with two consecutive epochs missing due centre. model. to a large block of time scheduled with Carrying on such large campaigns 2. Under the auspices of the "World an instrument wh ich was not suitable for requires good will, excellent organisation Astronomy Days", sponsored by ESA in our projecl. Again, the ESO schedule and communication among the astrono­ the context of the International Space matched as weil as possible the IUE mers involved in the pilot group, and a

27 broad consensus in the community on the References Peterson, B.M., Balonek, T.J., Barker, E.S., et importance of the project. We have been al., 1991, ApJ368, 119. Reichert, GA, Rodriguez-Pascual, P.M., AI­ fortunate in benefitting from the interest, Alloin, D., Santos-L1eo, M., Peterson, B.M., et loin, D., et al., 1994, ApJ425, 582. encouragement, and support of the staff al., 1995, A&A 293,293. Blandford, R.D., McKee, C.F., 1982, ApJ255, Robinson, A., 1994, in Reverberation Mapping of various observatories at which these 419. of the Broad-Line Region in Active Gatactic observations have been made. In addi­ Clavel J., Boksenberg A., Bromage G., et al., Nuctei, ed. P.M. Gondhalekar, K. Horne, tion to the tangible scientific return from 1990, MNRAS 246, 668. and B.M. Peterson, ASP Conference Se­ these programmes, we believe that the Clavel, J., Reicher!, GA, AIIoin, D., et al., ries, 69, 147. Stirpe, G.M., Winge, C., Altieri, B., et al., 1994, large-scale international collaborations 1991, A pJ 366, 64. Clavel, J., Wamsteker, w., Glass, 1., 1989, ApJ ApJ 425, 609. in the AGN field have greatly enhanced 337,236. Ulrich M.H., Boksenberg A., Bromage G., et the mutual interactions of the astrono­ Korista, K.T.,Alloin, D., Barr, P., etal, 1995, ApJ al., 1984, MNRAS206, 221. mers involved in the project, have led to a Suppt. Ser., April 15. much more efficient use of telescope Peterson, B.M., 1988, PASP 100, 18. time, and have resulted in a better Peterson, B.M., 1993, PASP105, 247. Peterson, B.M., 1994, in Reverberation Map- coordination of programmes, thus lead­ ping of the Broad-Line Region in Active ing to faster and unquestionable pro­ Gatactic Nuctei, ed. P.M. Gondhalekar, K. gress in our understanding of the AGN Horne, and B.M. Peterson, ASP conference D. Alloin, e-mail: phenomenon. Series, 69, 1. ALLOIN%[email protected]

On the Variability of Narrow-Line Seyfert 1 Galaxies

M.E. GIANNUZZ01 and G.M. STIRPE?

10ipartimento di Astronomia, Universita degli Studi di Bologna, Italy 20sservatorio Astronomico di Bologna, Italy

1. Introduction The distinction between type 1 and larger than that of the forbidden lines1 1 type 2 Seyferts is by no means clearcut. (FWHM ~ 1000 kms- ). Studies of NLS1 s Narrow-Iine Seyfert 1 (NLS1) galaxies Spectropolarimetry (e.g. Antonucci & have shown that the broad components are characterised by the relatively low Miller, 1985) has revealed that several of the lines have ratios similar to those of projected velocities of their line-emitting (though not necessarily all) Seyfert 2's 'normal' Seyfert 1's and, on average, nuclear gas. We describe a spectroscop­ contain BLRs which are hidden to con­ lower equivalent widths (Osterbrock & ic programme based on a search for ventional spectroscopy by obscuring Pogge, 1985, Goodrich, 1989); this last variability, which attempts to constrain material (possibly a dust torus). This has property, however, is the extension to low the causes of their difference with sparked a debate on the possibility that all FWHM of a trend observed throughout respect to other Seyfert 1 galaxies. Seyferts may be described within a the Seyfert 1 population. Some NLS1 Active galaxies which are classified as unified model, in which the orientation of galaxies present in their spectra high Seyferts (characterised by a Juminous the nuclear axis determines the aspect of ionisation iron lines like [FeVIIl )",5721, nucleus of stellar appearance, with a a source's spectrum, and therefore its )",6087 and [Fe Xl )",6375 (Osterb rock, non-stellar blue continuum and strong classification. Within this framework, 1985, Osterbrock & Pogge, 1985), in emission lines) are divided into two Seyfert 2 nuclei are viewed at large some cases with high intensity: these are categories according to the widths of inclination angles, and Seyfert 1's at properties common in Seyfert 1 galaxies, their lines: in Seyfert 2 galaxies, forbid­ medium and small ones. but quite rare in Seyfert 2's. NLS1 s den and permitted lines all have the same comprise approximately 10% of optically 1 width (- 1000 kms- ), while in Seyfert 1 selected Seyfert 1's, but a significantly galaxies the permitted fines have an 1.1 What is a narrow-Iine higher percentage - 16-50%) of soft X­ additional component of much greater Seyfert 1 galaxy? ray selected Seyfert 1 sampies (Steph­ 4 width (- 10 km S-1). The difference is ens, 1989, Puchnarewicz et al., 1992). attributed to the presence of both a broad The broad components of Seyfert 1's Boiler et al. (1995) report the observation line region (BLR) and a narrow line region display a great variety of profiles and with ROSAT of a sampie of NLS1 s, (NLR) in the nuclei of Seyfert 1's, while widths (e.g. Osterbrock & Shuder, 1982, finding that the objects in this class have only the latter is present, or visible, in Stirpe, 1990), and it is tempting to explain generally steeper soft X-ray continuum Seyfert 2's. The BLR is characterised by it on the basis of projection effects. In slopes than normal Seyfert 1's, and rapid higher densities, higher velocities of the particular, the so-called 'narrow-line Sey­ soft X-ray variability. gas which forms it, and by a smaller size fert 1 galaxies' (Osterbrock & Pogge, than the NLR: in fact, BLRs are so 1985) are at the lower end of the broad compact (<< 1 pc) that even in the closest line width distribution in the Seyfert 1 'It is important to realise that we are reterring to active galactic nuclei (AGN) they cannot class. While they are clearly distinct from objects whose maximum observed velocities tram be resolved spatially. The large velocities Seyfert 2's because of the different widths the BLR are low, not to objects in wh ich the broad component 01 the emission lines is very weak present in the BLR are generally attribut­ of permitted and forbidden Iines, and compared to the narrow component, but also very ed to the gravitational effects of a because of the presence of Fe 11 lines broad: the FWHM 01 the permitted line (broad + massive (> 107 M,,) accreting black hole, (which are not emitted by the NLR and are narraw component) can be similar in objects 01 these two types, and sometimes a low signal-to-noise ratio which is the prime cause of the nuclear therefore absent in Seyfert 2 spectra), the in a spectrum can mask a weak but very braad activity. width of their broad components is barely component, and cause an object to be misclassilied.

28 The question which we wish to ad­ fact, it is not even known whether NLS1 s chosen for the calibration is the same for dress is what causes NLS1 s to have such are optically variable, except for one the two spectra of each object, we could narrow lines: this can provide insight into case (NGC 4051, Peterson et al., 1985). therefore find a scale factor to correct for the more general problem posed by the Yet this information could be very useful in one of the two spectra, so making it great diversity present among the broad to constrain the hypotheses listed in the comparable to the other one. The calcula­ emission Iines of AGN. Some of the previous section on the nature of NLS1 s, tion of the correction parameters is possible answers are: as pointed out by Robinson (1995). A performed by a Fortran code (see van 1. NLS1 s are not intrinsically different lack of widespread variability, in fact, Groningen & Wanders, 1992) which from other Seyfert 1's, and the low would suggest the absence of broad line­ makes use of a chi-square minimum velocities in their lines are caused merely emitting gas very close to the central research procedure on the difference by projection effects. If, for example, the black hole, and would therefore indicate spectrum in the wavelength range includ­ BLR has a flattened configuration in that the BLR is located at higher distance ing the forbidden lines; together with a which the gas moves preferentially in the from the centre than in normal Seyfert scale factor and a shift in pixels, it also plane perpendicular to the axis of sym­ 1's, or that the inner and most responsive gives in outputthe FWHM (in pixels) ofthe metry (as in an accretion disk), our line of region of the BLR is obscured. If instead Gaussian with which one of the two sight towards NLS1 s should form a small variability is as common in NLS1s as in spectra may be convolved to best match angle with the axis itself. Seyfert 1's, this could imply a smaller the other. We found that this method 2. The main difference between NLS1 s central mass or an anisotropic kinematic gives better results, i.e. smaller residual and normal Seyfert 1's is the mass of the structure for the BLR. fluxes in the difference spectra in the central black hole, which is smaller in the Therefore, we have performed a region of the chosen forbidden lines, than former type of object, and therefore simple observational programme to de­ the method of direct evaluation of the causes the BLR gas to move at lower termine whether variability is a common forbidden line fluxes through a (e.g. velocities. characteristic of NLS1S. Gaussian) fitting of the line profiles, not 3. The broad line gas in NLS1 s moves only since allowance is made for a slight at lower velocities because it is on 2. Observations and Preliminary shift in wavelength, but also because average at larger distances from the Results there is no need to make hypotheses on black hole than in normal Seyfert 1's. In The programme consists in the obser­ the shape of the lines, which is often far this scenario, the BLRs of NLS1 s have a vation of a sampie of NLS1 galaxies at from being Gaussian. A problem with this larger emissivity-weighted radius than two epochs separated by about one year. method may exist, however, when the those of normal Seyfert 1'so For each object we obtained spectra NLR is spatially resolved and its exten­ 4. The inner region of the BLR, in covering the main optical lines, and sion is comparable with the projected slit which the gas moves at the highest compared the integrated line fluxes width: in this case, in fact, a different velocity, is hidden from our sight by measured for the two epochs. The results seeing effect in observations separated orientation effects: it is possible, within of this search for variability for our sampie in time may cause different portions of the the Seyfert unified model, that NLS1 s are were then compared with those of a narrow line fluxes to enter the instrument, objects seen at relatively large inclination larger existing data-base on 'normal' and therefore lead to errors in the angles, and that only the outer part of the Seyfert 1's, obtained with the same calibration. We are currently testing our BLR is observed. (relatively long) time scale. The sampie data against this source of error, by a we have chosen consists of 12 objects, quantitative evaluation of the spatial 1.2 Variability and is formed by all the NLS1s known in extension of the nuclei; however, the appearance of the bidimensional images A common characteristic of Seyfert 1 the literature with mv :0; 16.0 and intrinsic seems to exclude the presence of light nuclei is their strong variability: the UVI luminosity comparable to that of known losses caused by the seeing, since our optical continuum and emission lines variable Seyferts, and which were suited galactic nuclei are compact and the vary on time scales of a few days if not to our observing conditions. The obser­ observing conditions were quite good. An less. Normally the emission line varia­ vations were performed at the 1.52-m exception is represented by NGC 1365: tions lag those of the continuum by a few ESO telescope located at La Silla, Chile, its extended and composite nuclear days or weeks, indicating that the size of during two observing runs in early structure requires special attention in the the BLR is less than a few tenths of a October 1993 and late September 1994. analysis of the narrow lines. parsec. A great effort has been invested In both cases we used the same instru­ during recent years in the monitoring of mental configuration, covering the 3700­ To achieve an accurate calibration in Seyfert 1's (see Robinson, 1994 and 7500 A range at aresolution of - 1.9 AI the covered spectral range, we applied Alloin et al., 1995, and references quoted pxl. The S/N ratio of - 50 allows us to the internal calibration method separately in both reviews), in an attempt to un­ detect or exclude flux variations down to to the Ha part of the spectra, making use ravel the structure of the BLR through a level of - 10%. of the [S 11] A6717.0, A6731.3 Alines, and the technique of reverberation mapping The spectra were taken with standard to the Hß region, through the [0 111] (Blandford and McKee, 1982). Because procedures, and reduced making use of A4958.9, A5006.8 A lines. To actually of the large amounts of telescope time the standard IRAF reduction tasks. Be­ calculate the integrated flux for the required for adequate monitoring cam­ fore being able to compare the spectra strongest optical lines (mainly Ha and paigns (which typically consist of one obtained at different epochs, however, Hß), we first fitted and subtracted the observation every few days for several we performed a sort of 'internal' calibra­ continuum under the lines, and then months), care has always been taken to tion to correct for differential slit losses in evaluated the flux through an interactive select targets which were weil known to each couple of spectra, making use of the IRAF task. be highly variable, e.g. NGC 4151, NGC strong forbidden lines present in the The major sources of uncertainty on 5548, NGC 3783, Fairall9. All the targets Seyfert spectra: as mentioned previous­ the estimated line fluxes are due to the chosen for monitoring have broad com­ Iy, the forbidden lines are emitted by the process of internal calibration mentioned ponents of normal widths, so the results NLR, which is much larger than the BLR, above and to the evaluation (by hand) of obtained so far do not necessarily and therefore remain constant on time the fluxes themselves. Notice that, the generalise to the Seyfert 1 population as scales of decades. Imposing that the [0 111] lines being very intense compared a whole, and in particular to NLS1 s. In integrated flux of the forbidden lines to the [S 11] features, the calibration of the

29 Hex they were all Southern Seyfert 1's wh ich . -,----, were known when the programme was started (1979). The objects have been 15.0 - spectroscopically monitored at the ~30.0 - 1.52-m ESO telescope on long time scales, more precisely fram one year to 25.0 the next for about 15 years. 12.0 - To use this optical spectra data-set as ,....--. a comparison sampie, we first evaluated, I 20.0 - approximately with the same criteria used ~ 11 for our sampie, the Hl3 flux for each object 'I 9.0 vv" and each observation epoch (sometimes rn 15.0 calculating the mean value of more '"I observations taken a few days apart) Ei 10.0 II --- constructing an Hß light curve for each ü 6.0 tlD Seyfert 1 galaxy. We then calculated the .... I relative variation of the line flux for each Q) 5.0 NJ ~ 1-year interval. Since this time range is I much greaterthan the typical time scale of 0 1- I 0.0 ,.-< line and continuum variations, wh ich J- TI ranges from days to months (see § 1.2), ~'" --3.0 -- -0.8 we can assume that the annual relative variations are independent, being proba­ bly associated to different 'events' (bursts -6.0 - - -1.5 or declines of luminosity). We therefore constructed agiobai histogram, including all the annual relative variations (in -9.0 -2.2 absolute value) for all the objects, which _L ---L I I I 6600 6800 7000 5000 5200 in this wayforms a statistically meaningful sampie on variability data. A. (p,m) The data on the Hß line flux show that Figure 1. virtually all the monitored galaxies display significant variability, at least in some of HI3 region is usually more accurate than of a sampie of about 20 Seyfert 1 the time intervals covered. To determine that of the Ha region. galaxies, selected with no previous how the NLS1galaxies behave as aclass, We considered reliable only the meas­ knowledge of variability characteristics: compared to the 'normal' Seyfert 1 ured flux variations above 30', where 0' is the estimated uncertainty on the varia­ tion, finding that out of 12 objects, 9 show 19 - appreciable variations between the two epochs. In particular, two objects (IRAS 18 0345+0055 and NAß 0205+024) show 17 no variation and another one (Akn 564) displays a very weak decrease below our 16 threshold; four objects (ESO 012-G21, 15 Mkn 359, Mkn 9152 and NGC 1365) 14 underwent strong increases in luminosity (in the range 20-40%). We detected a 13 marginal variability in H 0707-495 and ~ 12 Mkn 1044 (in the range 10-20%), while () ~ 11 for three objects (IRAS 0444-052, Mkn .D 1126 and Mkn 896) we observed a clear ~ 10 variation only in one of the two main 0 ... 9 (lJ emission lines (Ha and Hß). .D In Figure 1 we plot the two spectra S 8 ;:l (1993 and 1994) of ESO 012-G21 and the z 7 difference spectra showing clearly the 6 line flux increase both in Ha and Hß. 5 3. Comparison with the Seyfert 1 4 Sampie and Discussion 3 The de Ruiter and Lub dataset with 2 -- which we compared our results consists

0 'Notice, however, that this objecl may have a very broad bul weak component in its emission lines (wh ich in ourspectra appears only in lhe Hllline), and therefore may have been misclassified as a NLS1 (see note 1). Figure 2.

30 population. we drew a histogram as close to the centre (as in 'normal' BLRs), The fact that variability is detected in described above for our results too: the and therefore that a small black-hole so many NLS1s is in any case a strong number of objects in the sampie is too mass or prajection effects are to be indication that these objects are not small to easily analyse the observed responsible for the low FWHM observed. radically different, in size and nature, variability fram a statistical point of view; In a relatively long monitoring, moreover, from normal Seyfert 1 galaxies. nonetheless, there is no apparent trend of it would be possible to measure the lag the NLS1s towards weak or absent between the line and continuum light Acknowledgements variability characteristics; on the contrary, curves to have a better estimate of the the objects for wh ich we clearly detected BLR size, and compare it to that of other We are grateful to Hans de Ruiter and a flux variation appear to distribute in the Seyferts. Jan Lub for allowing us to use their data in histogram in a similar way with respect to At this stage we can set, for the advance of pUblication. the comparison sampie, e.g. they can size of the observed BLR in NLS1s, an reach relative variations around 30-40%. appraximate upper limit of the order References In Figure 2 we plot the two histograms of the light-year, since we would not described above, which show no evi­ observe any significant variations if the Alloin, D., Santos-Ueo, M., Stirpe, G.M., and dence for a weak variability in the NLS1 line-emitting gas would be located tens Peterson, B.M. 1995, The Messenger, in sampie, especially when only the f1ux or hundreds of parsecs from the centre press. variations above 3cr are considered. (as happens in the NLR). The observed Antonucci, R.R.J., and Miller, J.S. 1985, Ap.J., 297,621. While the absence of variability, or a BLR could represent therefore the outer Blandlord, R.D., and McKee, C.F. 1982, Ap.J., significantly lower variability with respect part of a 'normal-sized' but partially 255,419. to 'normal' Seyfert 1's, would have had obscured BLR, or the entire BLR in a Boiler, Th., Brandt, WN., and Fink, H. 1995, strang implications on the interpretation type of object in wh ich for some reason Astr.Ap., in press. of the NLS1 s' spectrum, excluding that a there is no line-emitting gas in the inner Goodrich, R.W 1989, Ap.J., 342, 224. smaller central mass or projection effects parsec region. Osterbrock, D.E. 1985, PAS.P., 97, 25. Osterbrock, D.E., and Pogge, R.W 1985, could be responsible for the narrower We can notice, however, that the Ap.J., 297,166. lines, the presence ofvariability on a time annual relative variations measured in Osterbrock, DE, and Shuder, J.M. 1982, scale of one year does not entirely our comparison sampie appear to belong Ap.J.Suppl., 49,149. exclude the possibility (among the four to a long-term trend approximately in 30% Peterson, B.M., Crenshaw, D.M., and Meyers, KA 1985, Ap.J., 298, 283. Iisted in 1.1) that we are actually of the cases (estimated by taking the § Puchnarewicz, E.M., Mason, K.O., C6rdova, observing braad line emitting gas located events in which at least 4 points in the light FA, Kartje, J., Branduardi-Raymont, G., at relatively high distance from the centre curve show the same variation sign); Mittaz, J.P.D., Murdin, P.G., and Allington­ (which represents the whole or only the therefore, a common variability in NLS1 Smith, J. 1992, M.N.RAS., 256, 589. outer part of the BLR); this gas could in galaxies, as that displayed by our data Robinson, A. 1994, Proc. RAL Workshop on "Reverberation Mapping 01 the Broad-Line fact be insensitive to variations on short­ (- 75% of objects varied in luminosity Region in Active Galactic Nuclei", ASP Cont. time scales, but responsive to long-term between the two epochs), is more proba­ Se ries 69, 147. trends similar to the ones easily bly consistent with models in which a BLR Robinson, A. 1995, M.N.RAS., in press. recognisable in the Hß light curves of the with 'canonical' size produces smaller Stephens, S.A. 1989, AJ., 97, 10. de Ruiter and Lub data-set. To discrimi­ observed velocities of the line-emitting Stirpe, G.M. 1990, Astr.Ap.Suppl., 85,1049. Van Groningen, E., and Wanders, 1., 1992, nate between the competing models it gas, either for projection effects or for PAS.P., 104, 700. would be necessary to monitor a variable intrinsic reasons. We therefore tend to NLS1 (e.g. NGC 4051) with quite short favour these types of explanation, that time scales (days orweeks), such thatthe should be theoretically modelied to be presence of variability would surely imply able to compare in detail our predictions M.E. Giannuzzo, e-mail: that the line-emitting gas is located very with the observed line profiles. [email protected]

The DUO Programme: First Results of a Microlensing Investigation of the Galactic Disk and Bulge Conducted with the ESO Schmidt Telescope

C. ALARD 1, J. GUIBERT1 , 0. BIENAYME2, D. VALLS-GABAUD2, A. C. ROBIN3, A. TERZAN4, E. BERTlN5

1 Centre d'Analyse des Images and Observatoire de Paris (OEMIRM), France; 2 Observatoire de Strasbourg, France; 3 Observatoire de Besan90n, France; 4 CRAUObservatoire de Lyon, France; 5 Institut d'Astrophysique de Paris, France

Summary For this purpose, we used a first set of Some ten microlensing candidates, in­ We present the first results of a search Schmidt plates taken on La Silla fromApril cluding what appears to be an amplifi­ for microlensing amplifications towards to September, 1994, digitised with the cation by a double lens, could be present the Galactic centre region, aimed at MAMA microdensitometer, and analysed in the data produced by the reduction of investigating the populations of the disk with a software specially developed for half of the field. Thousands of variable and bulge in a wide field. highly crowded fields. stars are also evidenced by this survey.

31 about 12 million objeets on eaeh Sehmidt plate. u: 0,543803 t: 2,217504 As was done for the EROS projeet

2,2 f-- (Aubourg et al., 1993), digitisation of the photographie material is performed with the fast and accurate mierodensitometer MAMA (Maehine Automatique aMesurer pour I'Astronomie), designed and op­ erated by INSU (Institut National des Seiences de l'Univers/CNRS), and 10­ eated at Observatoire de Paris. The seientifie objeetives of this projeet include: - deteetion of low luminosity (if not dark) objeets through mierolensing am­ plifieation. The EROS and MACHO groups reported the deteetion of mierolensing eandidates towards the Large Magellanie Cloud (Aleoek et al., 1993,Aubourg et al., 1993), in the course of investigations -50 o 50 10e aiming at searehing for baryonie dark days matter in the Galaetie halo. Towards the region of the Galaetie 3.6 centre, mierolensed as weil as mierolens­ ing objeets ean apriori be loeated either u: 0.678425 t: 4,134663 in the bulge, or in the disk.lt is worth noting that our knowledge of the faint end of the 3.4 stellar luminosity funetion is presently rather paar, even in the solar vieinity. The OGLE and MACHO groups have ~ 3.2 al ready reported the deteetion of ::J 4-­ "'-../ eandidates in this direetion using CCD (J1 o deteetors (see, for instanee, Szymanski 3.0 -. etal., 1994and referenees therein, Aleoek lJ)* ., N et al., 1995a). The unique feature of the "'"f'"" DUO programme is its widefield whieh will 2.8 ;f; '. allow us to produee a large-seale map of the mierolensing optieal depth. This is partieularly important in orderto disentan­ 2.6 gle the eontributions of lenses situated in the disk and in the bulge, sinee their spatial distributions are very different (see, for example, the theoretieal maps -100 -50 o 50 100 ealeulated by Evans, 1994 and Kiraga, days 1994) - strueture of the bulge, and par­ Figure 1: Two examples ot microlensing candidatas so tar extracted trom the data. Data trom blue plates. The amplitications are close to 0.8 and 0.6 mag. respectively. At rest, star#a (top) is tieularly of the Bar, the eharaeteristies of tainter than star #b (bottom) byabout 1.4 mag, hence the larger error bars. For significance ot which still remain largely unknown at the the parameters, see Paczynski (1991): t the time it takes the relative positions otthe source and present time. lensing star to change by the apparent Einstein radius Re; u is the impact parameter (RIRe), - study of stellar populations in general, using the multi-eolour photome­ try whieh will be available for a large number of stars. This wealth of information will eon­ tween the Sun and the Galaetie centre. - short-period variables. tribute, together with the results obtained The area of one photographie plate (30 - long-period variables. by the OGLE and MACHO groups, to square degrees) covers a wide range of Several types of short- as weil as long­ improve our knowledge of the stellar galaetie coordinates, in the ease of the period variables are interesting distance populations and galaetie strueture in this DUO field : from bll =-4 to bll =-1 0, and indieators. As a eonsequenee, extensive region. from eil =°to flil =+6. monitoring of a large number of objeets is In this respeet, DUO is eomplementary expeeted to signifieantly improve our 1. The DUO Project and its of the MACHO and OGLE CCD projeets, knowledge of the 3-D distribution of Objectives whieh ean benefit, among others, from a stellar populations in the bulge, partieu­ denser time sampling, but over a more larly if the extinetion ean be estimated The Freneh DUO ("Disk Unseen restrieted area. from visible and/or infrared multiband Objeets") projeet (Alard, 1995a, Alard et In the region surveyed here, the stellar photometry. al., 1995) takes advantage of the large density is partieularly large, but the Similar "by-produets" of mierolensing field of the ESO Sehmidt teleseope to extinetion is moderate and relatively experiments are being obtained in the investigate the stellar populations be- homogeneous, allowing monitoring of EROS, MACHO, and OGLE projeets

32 5.4 C type RR Iyrae 2.' Cantaet Binary Cantocl Binary " SX Phaenix 5.2 I~. I:~ 2.' 50.... ." '('I, -1:... '('I' 4:8 .... J' ..... J' * •••• ~ •• :t •• '"~ \io~ 1...,.'" ...... 4.6~p~p 2. ...; ~ 4.4 .., " 4.2 Period: 0.34062 days Period: 0.37430 days Period: 0.47665 days Period: 0.15193 days ,,~~.• ~""'-"';•."","-'--'--'-:';:-,-,~-f'.':-'-".-...... ::'2 • 0.0 0.5 1.0 1.5 2.0 0.0 •., .., 2. o. 0' '.0 .., 2.• ,., ,., A type RR Iyrae SX Phoenix Beta Lyrae eelipsing variable ..• ,. A type RR Iyroe 2.' .", .", :\..... (\.. ,. " ... ~... :\...~.,.:\ •• .., .,;,,. 4.0 Period: 0.60642 doys ,.. Period: 0.08568 days 2.. Period: 0.43798 days Period: 0.45645 days 00 .., ... .., ,. ... 0.' ,. 2.• 0.0 •., 2.•

,., A type RR Iyrae A type RR Iyrae Cantaet Binary '" Cantaet Binary '.0 .. .,~. .~. Ja .... ) ...... • \ '. ,., ~ '" «... « lot:d;..t \\ 111 \\ 111 " "...... •• "...... I 2.• Period: 0.59168 days ,.. Period: 0.54424 days Period: 0.35062 days 2.• Period: 0.61460 days

0.• 0.' ... ,., 2.• •.. 0.' ,.. .., 2...... , .., 2.0 ... .., ... 2.•

'.0 B type RR Iyrae U Contoct Binary J.Q Algol type eclipsing variable Algol type eelipsing variable " ....'~.,...... '~., ...... ,.. .. ••. I u ·f• ". 2.• • '" Period: 0.55383 days Period: 0.44998 days ,., Period: 0.65629 days ,., Period: 3.46020 days 0.• 0.' ... 2.• 0,0 0.5 1.0 1.5 2,0 ... 0.' ..• 2•• Figure 2: Some examples of the 10,000-15,000 new short-period variables expected from the DUO project. The variability type, derived trom the shape and characteristics of the light curve, is indicated at the top of each box. The phase appears in abscissa, and the amplitude (in mag.) in ordinate. The light curve is repeated once for c1arity.

(e.g.: Udalski et al., 1994a, Grison et al., in this region, are within the reach of the GGD frames have been obtained at La 1995, Beaulieu et al., 1995). project. Silla with the ESO/Danish 1.54-m tele­ The digitisation of one ESO Schmidt scope. 2. Observations and Scanning plate, performed with a 10 ~m (0.67 At the time of writing (April, 1995), half arcsec) step, yields aseries of FITS of the field has been entirely reduced, The plate material for the 1994 images requiring a total storage space of resulting in 8 million light curves. For each Campaign is composed of about 200 1.6 Gbyte. star, the stored information comprises the plates, mainly IIlaJ and IIlaF, taken with coordinates of the object on the reference the classical GG385 and RG630 filters, plate, and, fortheotherplates, threebytes respectively, thus giving access to the 3. Reduction and Preliminary describing the magnitude and a confi­ blue and red parts of the spectrum. A few Results dence indicator. Therefore, for this zone, plates have also been obtained in other the requested storage space is 6 Gbytes. bands, namely with the combinations In the DUO field, the stellar density, as From this significant amount of data, a IIlaJ/UG1, IllaF/GG495, IVN/RG715, for detectable on Schmidt plates, turns out to first set of microlensing candidates, more detailed characterisation of the range from 200 (in the north-west corner), among them an object with a light curve stars. to 100 (south-east corner) stars per suggesting a binary lens, has al ready The plates were taken over a six­ square millimetre-i.e. persquare arcmin. been detected. month period around June 1994, most As a result, a total number of about 12 A high number of periodic variable ob­ often two IliaJ and one IliaF plates per million stars can be monitored. jects has also been discovered, essential­ dark night, all three with 20 minutes A special software has been designed Iy eclipsing binaries and RR-Lyrae stars of expOsure. and developed by one of us (Alard, 1995b) various types. A total number of 10,000­ The limiting magnitude is close to 21 for optimal detection and photometry of 15,000 such objects is expected for the and 19 in the blue and red colours, this huge number of objects in a rather whole DUO field. Long-period variables, respectively. In view of the time sampling crowded field. The resolution achieved is for instance Miras, and irregular red var­ obtained in the blue band, microlensing better than 0.1 mag. for most of the iables, are also numerous in the region. events of short durations as weil as short­ monitored stars, and often better than These first results are hereafter pre­ period variables, known to be numerous 0.05 mag. For the magnitude calibration, sented in some detail.

33 4. Microlensing Candidates stars discovered in the course of our events, and for the study of the long­ survey. The most numerous are contact period variables. This is the reason why The theoretical microlensing magni­ binaries with periods smaller than one we have applied for a second run at the tude variation has been fitted to the day, followed by Algol-type objects and ESO Schmidt: lila plates and Kodak4415 observed light curves after a preselection Beta-Lyr systems, this ranking being Tech Pan films will be taken on La Silla made among the time series, on the basis consistent with the results obtained bythe from May to August 1995. of consecutive three-cr deviations with OGLE group (Udalski et al., 1994a) on respect to the minimum. The achroma­ 116 eclipsing binaries discovered in the Acknowledgements tism of the amplifications can be checked centre of the Baade's Window. The large only for stars which, when not amplified, number of eclipsing objects which will be It is a great pleasure to thank B. are sufficiently above the limiting magni­ produced by the DUO survey is expected Paczynski for fruitful discussions and tude of the red plates. The resolution is to make possible statistical investigations valuable suggestions. We also thank G. also better for the IIlaJ plates. of this population which has been only and O. Pizarro for the set of excellent From the reduction of the first half of poorly studied up to the present time. Schmidt plates taken at the ESO Schmidt the field so far performed, some ten RR-Lyrae stars represent about 20% telescope on La Silla, and the MAMA events appear to be reliable candidates of the variable objects so far detected in team for support during the scanning of for microlensing amplifications. Two ex­ the DUO field. These can be used to map this photographic material. amples, chosen among these, are shown the interstellar reddening, and are also in Figure 1. good distance indicators. They are Among the candidates, an object with therefore invaluable in investigations of References an unusuallight curve has been detected the structure of the Galactic bulge. The with three consecutive peaks within 7 latter issue is of prime importance, in Alard, C., 1995a "Workshop on the Future 01 days. This behaviour is quite surprising particular, for the study of the Bar. Microlensing Surveys", Livermore, Ca, Jan­ uary 13-15, 1995. for an intrinsically variable object, and the Although the existence ofthe Bar appears Alard, C., 1995b, in preparation. most likely explanation seems to be an to be weil established by now, its Alard, C., Mao, S., Guibert, J. Valls-Gabaud, amplification by a double lens (Alard et orientation and axis ratio are still poorly D. 1995, in preparation. al., 1995). The possibility of observing known. Alcock, C. etai., 1995a, Astrophys. J., in press. microlensing by multiple lens, anticipat­ This component of the galactic struc­ Alcock, C. et al., 1995b, Astrophys. J.. in preparation. ed by Mao and Paczynski (1991), has ture is receiving special attention from the Alcock, C. et al., 1993, Nature, 365, 621. al ready been put forward by the OGLE groups involved in microlensing projects Aubourg, E. et al., 1993, Nature, 365, 623. group (Udalski et al., 1994b) on candi­ (see, e.g. Kiraga, 1994 and references Beaulieu, J. P. et al., 1995, Astron. Astrophys., date OGLE #7, and later confirmed therein, Stanek, 1995 ,Zhao et al., 1995, in press. independently by the MACHO collabora­ and references therein). The optical Della Valle, M., 1994, Astron. Astrophys., 287, L31. tion (Alcock et al. , 1995b). depth to gravitational microlensing in the Evans, N. w., 1994, Astrophys. J., 421, L31. The confirmation of the suspected direction of the Galactic bulge appears to Grison, P. et al., 1995, Astron. Astrophys. candidates, with characteristic durations be in excess by a factor of 5-10 with Suppt., 109, 447. ranging from 3 to more than 60 days, will respect to the previously expected values Kiraga, M., 1994, Acta Astron., 44, 241. require scanning and reducing the whole (see, e.g. Evans 1994 and references Mao, S. Paczynski, B., 1991, Astrophys. J., stack of plates, including those taken in therein). This could be due to bulge-bulge 374, L37. Paczynski, B., 1991, Astrophys. J., 371, L63.. the U band, in particular to eliminate the gravitational amplifications, the Bar play­ Stanek, K. Z., 1995, Astrophys. J., 441, L29. possibilitythatthese events are produced ing a major role in this process if oriented Szymanski, M. et al., 1994, Acta Astron., 44, by dwarf novae (Della Valle 1994). towards the Sun as proposed by several 387. authors. Udalski, A. et al., 1994a, ActaAstron. , 44, 317. 5. Variable Stars Udalski, A. et al., 1994b, Astrophys. J., 436, L103. 6. The Near Future Zhao, H., Spergel, D. N., Rich, R. M., 1995, Figure 2 displays a set of short-period Astrophys. J., 440, L13. variables which are representative of the The reduction of the second half of the variety of the new interesting objects field is in progress. Additional observa­ detected in the DUO field. tions will be necessary, especially to Eclipsing binaries represent the improve the time base line, as weil as to J. Guibert dominant population among the variable increase the statistics of microlensing e-mail: [email protected].

The Variation of Atmospheric Extinction at La Silla G. BURKI, F. RUFENER, M. BURNET, C. RICHARD, A. BLECHA, P BRATSCHI, Geneva Observatory, Sauverny, Switzerland

1. Introduction Geneva System (Golay, 1980; Rufener, Earth atmospheric extinction in the optical 1988) have been obtained by using domain has been collected. A first A total of 248,000 stellar photometric successively two telescopes (40 cm and analysis of the atmospheric extinction measurements have been obtained from 70 cm), two photo-electric photometers variations was published by Rufener the Swiss station at La Silla from July 1977 and one CCD camera, from two different (1986, hereafter Paper I) for the period to August 1994, during about 4400 nights locations on the site of La Silla. from November 1975 to March 1985. The of photometric quality. These seven­ A very homogeneous set of photomet­ annual and long-term variations were colour photometric measurements in the ric data, and, in consequence, of data on described as weil as the effect of the 34 1980 1985 1990 1995 efficient absorbant when transformed into sulphurie acid by photochemical effect in 0.4 the presence ofwatervapour (Mroz et al., 1983; McCormiek & Swissler, 1983). The Pinatubo in the Philippines (Iati­ tude +15°) had also two main eruptions: 0.3 : .. the first of four pre-paroxysmal vertical .' : :::. i :.: ~,"J,.: . eruptions took place on June 12, 1991 >- . 0.. ,," i .~ : .: 0. : ;. ::"o"i-~: . (HJD 2448420), and the main eruption '-- ~ i~~ ~ 0.2 • I • '11 :. .;; .:, " ';, ;. -i, r1'i!il:. ' :. on June 15, 1991 (HJD 2448423). The estimation of the amount of S02 emitted ;t'M'~:';~M')::;\~*,'·:I~t~i.'ti: ~~,I~:- .:f,,::~~\}, .. ~1i~~~.~ ~ to the atmosphere is about 20 mega­ 0"' ; (. . • .' .".. ".1 1M.' l . . . . ". tons (Bluth et al., 1992; Pallister et al., 0.1 I'..ff· ;." ) 4 •• I ..' I.: .' ..! " 1992). As shown by Figure 1: (i) these two volcanoes ean be considered as the o ~-'-_-'--_.l...-_L.----,_--,-I_-1-_-1- ---'_--'-_----'-_--'-_.L...J cause for the long-term variations of the 4000 6000 8000 extinetion at La Silla du ring the past 17 HJD-2440000 years; (ii) the increase of the extinction at La Silla was very sudden, roughly 150 Figure 1: Variations of the extinction coefficient in the V filter. days (EI Chichon) and 100 days (the Pinatubo) after the eruptions; (iii) the effect from the Pinatubo was much voleanie aerosols injeeted in the Earth tinetion during the photometrie nights of strongerthan the onefrom EI Chichon; (iv) atmosphere by the eruption of the EI the past 17 years at La Silla are presented the decantation of the voleanic aerosols Chiehon voleano. in Figure 1 for the filter V. The extinction in the atmosphere is very slow, lasting This problem has been revisited on variations ean be deeomposed into: long­ at least 1000 days, perhaps even 1300 the basis of the new photometrie data term variations due to voleano aerosols, a days. (Burki et al. , 1995, hereafter Paper 11) mean annual variation and short-term Our data permit to follow the evolution beeause of: (i) the time period eovered by variations due to meteorologieal aero­ of the extinction law due to the volcano the observations, whieh is now 17 years; sols. aerosols, du ring some hundreds of days (ii) the atmospherie extinetion whieh has following the eruptions. Adopting a law of now been evaluated for all the photomet­ 2. 1 Long-term extinction the standard form k(A) - A"", we obtain rie nights (and not only for the so-ealled variations due to volcanoes that: (i) the aerosols from the two MD nights (see Paper I) representing voleanoes were very different at their about only 20% of the photometrie Two voleanoes are weil known to have origin (or, more precisely, when they were nights); (iii) the effeet of another voleano affeeted the Earth atmosphere's trans­ deteeted from La Silla) ,the aerosols from eruption, that of the Pinatubo, whieh can pareney, in particular at La Silla, during the Pinatubo having produeed a flater, or be analysed. the past years: gray, extinetion law; (ii) the "voleanie EI Chichon in Mexico (Iatitude +1r) aerosols" produee very different extine­ 2, The Variations of the had two eruptions, in March 23 and April bon laws than the "meteorological aero­ Extinction 4, 1982(HJD2445051 and2445064).The sols" (see Seetion 2.2); (iii) the evolution towards an inerease of with time, stratospherie loads due to EI Chichon is u p The general eharaeteristies of the estimated to reach about 8 megatons of observed in the eases of the two volea­ variations of the mean atmospherie ex- sulphur dioxide S02' a radiatively very noes, eould mean that the voleano

1 0.2 ..--. : : ...... > '. . 0.8 , . ,'. . ~ ...... ,...... 0 ' .. Cd 0.6 .-< >' (l) ~ 0.1 ~ - ~ 0.4 0 () ....,0 ;::l 0.2 o -< 0

III V VII IX XI I III V 0 20 40 60 Month T[days] Figure 2: The annual variation ofthe extinction coefficient in the V filter. Figure 3: Analysis of the series of the extinction coefficient determina­ tions in the V filter by autocorrelation.

35 aerosols have a size increasing with time, when the atmosphere is, for instanee, eharaeteristie time for these variations, until they fall on the ground. polluted by voleanie outbreaks or forest restrieted to the nights of photometrie From the theoretical calculations ofthe fires. When estimations eoneern speeifi­ quality, va ries from 2-3 hours to several extinetion law (van der Hulst, 1952), we eally the meteorologieal haze, the values years. The amplitude of these variations for a are less dispersed, between -1.25 derive an inerease of the typieal aerosol p is large: during the 4400 nights of our radius from 0.4 to 0.7 ~m from HJO and -1.45, and are independent of the photometrie aetivity, the proportion of the 2445170 to 2446030 (EI Chiehon), and haze density. light diffused or absorbed by the Earth from 0.5 to 0.8 ~m from HJO 2448670 to The eomponent of the extinetion atmosphere has varied, in the V band, 2449260 (the Pinatubo). Thus, thetypieal analysed in this Seetion is related only to from 9% (minimum extinetion) to 33% radius was a little larger for the Pinatubo the haze present during the nights of very (beginning of the period affeeted by the aerosols, and, in the ease of both good and stable, photometrie, atmos­ aerosols from the Pinatubo). voleanoes, we note an inerease of the pherie transpareney. This is an indieation In the ease of ground-based measure­ typieal aerosol radius with time. These that the distribution of the radii of the ments, the eorreet estimation of the fluxes eonelusions are in agreement with the particles forming the haze (mainly "outside the atmosphere" must absolutely results obtained by Hofmann & Rosen eonstituted of water droplets) is quite be done by using the atmospherie extine­ (1983, 1987), Knollenberg & Huffman stable during these nights and has its tion value for the site and at the time the (1983), Oberbeek et al. (1983) in the ease mode near the value 0.3 ~m. measure has been performed. In the of EI Chiehon and Valero & Pilewskie eases of stable photometrie nights, the (1952) in the ease of the Pinatubo. 2.3 Short-term variations of the mean value for the night ean be used. An extinction estimation based on the extinetion values 2.2 The annual variation of the of other nights will give eorreet results only extinction Is there any typical period of several by chance. days, or weeks, during whieh the extine­ The annual variation of the extinetion tion does not vary eonsiderably? Or, in References is mainly due to the variation of the other words, is the extinetion in a given Angsträm, A: 1961, Tellus 13,214. meteorologieal aerosol content in the night eorrelated to the extinetion of the previous and/or the following night(s)? Bluth, G.J.S., Doiron, S.o., Schnetzler, C.C., atmosphere above the observing site. Krueger, AJ., Walter, L.S.: 1992, Geophys. This variation ean be examined by Based on our large material, it is Res. Lett. 19,151. removing the minimum mean extinetion possible to give a elear answer to this Burki, G., Rufener, F., Burnet, M., Richard, C., values and the extinetion due to voleanie question, for the La Silla site, sinee the Blecha, A., Bratschi, P.: 1995, A&A in press (Paper 11). aerosols (Seetion 2.1) to the global data. extinetion has been determined during almost all the photometrie nights during Golay, M.: 1980, Vistas in Astron. 24, 141. In addition, forthe purpose ofthis Seetion, Hofmann, D.J., Rosen, J.M.: 1983, Geophys. aperiod of about 17 years. The best the data from the periods of two years Res. Let. 10, 313. following the eruptions of the two vol­ mathematieal method to analyse these Hofmann, D.J., Rosen, J.M.: 1987, J. Geo­ eanoes have been exeluded. The result­ data is to use an autoeorrelation teeh­ phys. Res. 92, 9825. Knollenberg, R.G., Huffman, D.: 1983, Geo­ ing extinetion values in V have been nique. The results, presented on Figure phys. Res. Let. 10, 1025. plotted in Figure 2 versus the date in the 3, are based on the k(V) values ob­ tained during the years 1985 to 1990 McCormiek, M.P., Swissler, T.J.: 1983, Geo­ year. We see that: (i) the minimum value phys. Res. Let. 10,877. of the extinction is relatively stable only (HJO 2446066-2448257), in order Mroz, E.J., Mason, A.S., Sedlacek, W.A.: throughout the year; (ii) the maximum not to be affeeted by the effeets from 1983, Geophys. Res. Let. 10, 873. values are reaehed during the first part of the voleanoes. An important autoeorre­ Oberbeek, V.R., Danielsen, E.F., Snetsinger, K.G., Ferry, G.v., Fong, w., Hayes, D.M.: the year, i.e. during the southern summer. lation appears for a time T ::; 5 days, shown up by a deerease of the funetion 1983, Geophys. Res. Let. 10, 1021. The extinetion law for these meteoro­ Pallister, J.S., Hoblitt, R.P., Reyes, AG.: 1992, logieal aerosols has been determined in from 0 to 5 days: the autocorrelation is Nature 356, 426. Paper 11 (the effeet from the voleanoes very high (0.90) for T = 1 day, and Rufener, F.: 1988, Catalogue of Stars Meas­ 139 ured in the Geneva Observatory Photomet­ having been removed): k (I..) = b 1..- . decreases until T ;:: 5 days. Conse­ P p rie System, Observatoire de Geneve. Thevaluea =-1.39(theAngströmfaetor) quently, the global tendeney is to have • p Rufener, F.: 1986, A&A 165, 275 (Paper I). ~ IS remarkably stable and weil defined at La series of a few (n 5) nights with Valero, F.P.J., Pilewskie, P. : 1992, Geophys. Silla. It is known (e.g. Wempe, 1947; similar extinction values. Res. Let. 19, 163. Angström, 1961) that a varies within a vanderHulst, H.C.: 1952, TheAtmospheresof p small range and that, under global, the Earth and Planets, Ed. G.P. Kuiper, Univ. of Chieago Press, Chieago. average eonditions, at a large variety of 3. Conclusion Wempe, J.: 1947, Astron. Nachr. 275, 1. loeations on the Earth, it has a value elose to -1.3 ±0.2, being seldom above -0.5 or The variation of the atmospherie below -1.6. The values of a larger than p extinetion is essentially due to the mete­ G. Burki -0.5 are eneountered under eonditions orologieal and/or volcanie aerosols. The e-mail: OBS.UNIGE.CH

36 OTHER ASTRONOMICAL NEWS

FITLYMAN: AMidas Tool tor the Analysis ot Absorption Spectra

A. FONTA NA, Osservatorio Astronomico di Roma, Italy P BALLESTER, ESO-Garching

Introduction makes it possible to independently derive these lines, it is often necessary to test the column density N and the Doppler different configurations (i.e. number of The introduction of echelle spectro­ width b from li ne profile fitting, rather than components and constraints on them) graphs on 4-m telescopes has allowed from the curve of growth analysis. Typical before a satisfying fit is achieved. In the the observation of faint objects (m < 18) applications of these techniques are the analysis of the Lyman-a forest, on the with high resolution (R= AI LlA > 20,000). study of the Galactic clouds, which trace contrary, hundreds of independent lines In the study of the interstellar and the gas content and the metallicity inside are fitted, wh ich must be identified by the intergalactic medium, where the absorp­ the Galaxy, and of the absorption systems user himself. In any case, a significant tion lines from tenuous and relatively cold found in the spectra of high redshift fraction of the time required for this (T:::; 5 x 104 K) clouds are sought in the quasars, wh ich trace the evolution of analysis is spent in user interaction rather spectrum of bright background sources, primordial structures. than in computing time. this resolution allows the observation of The fit of heavy element systems may Three different approaches to this narrow lines with an instrumental FWHM require the simultaneous fit of different problem are known: (a) interactive lower than their intrinsic width. This lines spread all over the spectrum. For construction of a Voigt profile until a

5B29.1-r<;-----~:_.:

5748 5750.4

FrT/\,:rt1AU , ••n Int.e,.IIC\.JvII procr (ar "he (ILW.,. ot ebl5Orpt.lon IiI"loM In h1ch ""011,1\.,100 IIPIICt....-. It...... _-drl...... Int.llrl'eo- La .J100l t.M 11 ••" \.0 PeC1F'1J Lhe Jines to bc rnt.ed. and dhpJ.W' the dote on the dcroltlt rrorhlc: wlndow. Je. pe ... ro,... ,talMI du-.que,.••Inl_l.tet.lon 1,111101: toh. H.IHUtT pllek CII d J0"'" et. CERH. -) LOoIldJr'8 & Jnil1.YJlztnQ duU .•• n ce--plele t....ce Cl .11 t.he 'tL. "rfo~d I...tored In -) Mllil pl.ai~.. .s1,l1t.od "oblcu. ond con bo lI.od \.0 rO-Ito,.t I'Un1N~.Uon. ror .. ruH IIxph,neuon cf the ~1,1 runctJ.onc~ pl••u ,..f.,. LO the -) ..don. ..nu.l.

F"lTl'L'rtWl [ .c] [out.t.b] [10ll (union) [1".....h(V...H»)

--) Sf,MM;ll~ rro. 5200.05 to 7021.63 ( 18212 poi"l.w) e~(' • ep.ctrl.ll'l 1n tab1. FOT'Tllat. r.bl. 1l1U.t. cOflt.ln 4 col...n.: :WI1Y[. :STOCV. :HODtrUD( .,..d •••••••••••• flllmAN I'IEJ«J •••••••••••• l..-Wl+t. It. can be or."t.ed u.u'C t.hlt COtNERTI'lY'HAH Oa-4lnd. OeF'oult 1* the SPECTnB k rd.

PARAtETt.RS Cff.R:ATIUNS GR~ICS

eH)... (S)TAf'IlAF4) I1INIt1. deflrw (w) trldow (I)llU"61l. ecHt (P)ar..,. plot (B)Yckwardli' lolild rrOlll 10(&) .eil l (L) i.tl.• plOl (F)OrwMf"d 4<11 t (H) 1nu1 t (C)urYOr b.l

rolr( and QU;k n lhb mco 10 CCCCS3 help files

History.. Search... Nows Pli'"

Figure 1: Example ofa FITLYMAN session. The Main Menu is shown in the bottom left window. The graphie window overit displays a fit performed On a complex metal system at z =2. 768. Lines fitted are, from the top, A11670, CIV1550, CIV1548, 5i1/1526. The CIV doublet has been fitted with three components, lower ionisation elements with five. Data from Giallongo et al., 1993, and J. Wampler.

37 Here, r °is the classical radius of the electron, c is the speed of light, 'Ac, fand rare the three parameters associated ,,"0 ~ • + with the observed transition (Le. rest 1+++++-t=fl- t frame central wavelength, oscillator strength and damping coefficient, re­ ...0 ttt'i~ ...... +- spectively). In general, the Doppler broadening of t t + a li ne is due either to the intrinsic thermal 30 + broadening b = (2kT/m)O.5 or to a +t+ • K turbulent motion of the gas bT = herr' y t where aTis the inner velocity dispersion. 20 f 0+.+ ... • + When transitions from different ions are observed, it is possible to constrain the + +~ two cases: for the thermal component of 10 1 __ 1 ... 1 ... "'"0 the broadening, indeed, the b-values of ..... C ...... J ) different atoms scale with the masses as Figure 2: Results of a set of simulated fits. Squares represent the mode value (i.e. the most bion11bion2 = (mion2/mion1 )0.5, while in the probable value) of the median value of the b, N va lues obtained by the fitting proeedure, and the turbulent case they remain essentially error bars show the region where 68% of the points are found, representing an estimate of the equal. To allow for a complete analysis of observed seatter. Lozenges are the "true" parameters of the line. the physical conditions of the absorbing clouds, both types of b-values can be accounted for in computing the line satisfactory solution has been achieved Main Features profile: each line is thus parameterised by (Mar & Bailey, 1995) ; (b) construction of four different parameters ('Ac, N, bK, bT), and the Voigt profile is computed using a a photo-ionisation model, from which the We describe here the most important (b~ relevant absorption lines are computed features of FITLYMAN: for a more b-value b = + b; )0). This unique (e.g. using the Midas context Cloud) and complete description of the commands feature, combined with the possibility of finally compared with the data (Wampler and algorithms please refer to the ESO constraining the different parameters in a et al., 1995); (c) fit of a Voigt profile by Midas User's Manual. flexible way (see below), allows for the numerical minimisation of X2 (Carswell et • Each absorption line is modelIed by a simultaneous fitting of different ionic al. , 1991, Giallongo et al., 1993). Voigt profile, whose parameters are the transitions to yield meaningful values for Even though the two former methods central wavelength Ac, the column density both the turbulent and thermal broaden­ have the advantage of being more N and the Doppler broadening b. The line ing. intuitive and easier to implement, the profile f('A) for a spectrum 10(,1.) is I('A) = • An extensive set of atomic param­ latter approach is usually preferred, since fo(A)e-r(!I.) where the optical depth is eters is supplied: it allows for the fitting of it is more objective, testable through several ionic transitions found in absorp­ , N1i,fJltcl., lOK H( ) H( )' h tion spectra, and can be easily expanded numerical simulations, provides an 1\11.) = 1>021112 a, u. a,u Ist e accurate estimate of the errors, and is by the user him/herself. usually faster. Voigt function, defined as: • It is possible to simultaneously fit While the reduction of echelle spectra H(a u)=j- (·'-tl." lines from different ions, or from different to 1D calibrated spectra may be , -<.0:> (f2 +(11_.\')2 transitions of the same ion, specifying accomplished with several standard whose parameters a and u are given by: several links (relations) among them astronomical packages (such as the which reflect the physical condition in the MIDAS Echelle context), no similar tool u = (A,-Alt' cloud responsible for the absorption. The exists for the fit of their absorption hA, )21112 lines can be forced to share the same features. The lack of a software tool properly tailored for all the different aspects of this problem has moved us to the realisation of FITLYMAN, a new package available in the Lyman context of the Midas 94NOV release, that we present in this paper. This code is the evolution of a former version, developed at the Osservatorio Astronomico di Roma (Giallongo et al., 1993). It allows the fit of Voigt profiles to a large number of absorption lines in a normalised spectrum through the numer­ ical minimisation of X2 . Its main features are flexibility in specifying the line . .-_.. configuration, a fast user interface, and '. . 115 full documentation on the fit as performed. :Z~ Special care has been taken to 1:'. "7' 17' maximise the efficiency of the user operations: FITLYMAN provides recur­ Figure 3: Values of Dopplerparameter band eolumn density N obtained by the fitting proeedure sivity and allows recovery of previous for 1000 fines ofb = 30 and log N = 13.8 (Ieft) and for 1000 fines ofb = 20 and log N = 15 (right). configurations. Smal/ boxes show the respeetive line profile

38 minimisation. Furthermore, a batch procedure (BATCH/LYMAN) allows for the automatic, non-interactive fit of a 1~!G1-- - 00. :iI whole spectrum using the configurations stored in the "log" tables. • The FITLYMAN User Interface has been designed to satisfy two different ­ n and almost opposite - requirements: it r'l! !!I! I' I must allow the user to handle and modify r ' l t. all the parameters needed for the J' : minimisation, reducing the requested ,J 'I I '1 time as much as possible. To fulfill the :j.rJ 't. Midas requirements about compatibility 00 with any kind of terminal and environ­ 10 20 ment, we have designed a menu-driven User Interface which uses both the standard 1/0 devices (keyboard + screen) and the standard Midas graphics window. A Main Menu introduces the user to all the functions available in the programme, such as modifying the input parameters, velocity relative to the earth (redshift), complex situations. The original MINUIT loading/saving previous configurations, aM/or the same column density N, and/ package can also be used in interactive modifying the set-up options of the or the same b value, and/or the same mode: since this option is not compatible programme, or accessing the graphics temperature T by properly scaling (bK) with the Midas command language, it has device. The parameters which specify a ~ith the ion masses. Any combination of been disabled in FITLYMAN. Neverthe­ li ne configuration may be entered or lines and links is allowed, the only limit less, all its relevant functions can be modified either by answering a sequence being the maximum number of indepen­ invoked from the User Interface: this of questions that guide the user to the dent parameters which the programme solution also eliminates the necessity of correct sequence of operations (to be can handle (currently 1OO).ln general, the learningthe MINUITcommand language. preferred when new lines are defined), or number of free parameters depends on MINUIT may perform an accurate error by directly accessing their values (to ~he kind and number of links imposed: for analysis, by computing either the modify a few parameters). Instance, when the lines are forced to covariance matrix and/or more detailed The input spectrum and the fitted ~ave the same absorption redshift, there statistics, as non parabolic errors orscans profile are plotted on the active graphics IS only one actual free parameter (Le. the and contour levels in the X2 space. window. Optionally, the standard devia­ redshift), not one for any individual central • In principle, the results of each tion and/orthe residuals may be plotted as wavelength. minimisation depend on several parame­ weil. The cursor in the graphics window . The absolute freedom in specifying ters, which specify: (a) the line configura­ may be used to enter the wavelengths links with these rules is a powerful tool tion, (b) the spectrum regions where the needed to specify some parameters when the ionisation and metalIicity of a X2 is computed and (c) the MINUIT (e.g. "first guess" central wavelength or system is investigated, and alsoto correct commands used. To keep track of these the intervals over which the X2 is !or possible systematic errors: for parameters, each minimisation is associ­ computed). Instance, two lines far away in the ated with a unique identifier, whose value An example of a working session with spectrum (e.g. Si 111260 and Si 111526) is stored in the output table. At the same FITLYMAN is shown in Figure 1. rnight be forced to have the same band N time, the full initial configuration, together with the same identifier, is stored in three but different redshifts, to allow for some Performance and Examples wavelength calibration uncertainties. "log" tables. Later, these tables can be • The input data have to be stored into either consulted for the interpretation of To investigate the accuracy and aMidas table, containing wavelength, the results, or used to re-start the stability of the results obtained with normalised flux, standard deviation and instrumental FWHM for each pixel. A tool for the conversion of the usual image data 20 to this format is supplied. Both pixel size e- and resolution can vary along the spectrum. Gaps may exist in the ~pectrum, provided that they do not fall in s.. ~ fitted regions. • The numerical minimisation is at performed using the MINUIT package, - developed at CERN. It allows for the simultaneous minimisation of up to 100 2.<:> ~ e-~ !ree parameters (equivalent to at least 33 >. e- Independent absorption lines). The user -- can choose among different minimisation algorithms or strategies. For instance, '-s.~"~_ ~~s.~,.:=L-. -.~~,"",,",""""' 1 C> -.... __• ...... J ... , ~ome parameters may be fixed at the first .. 1 .... 10=0..,..... Iteration, to find a reasonable "first guess" Figure 5: Values ofDopplerparameter band column density N obtained by the fitting procedure solution, and then released to achie'"e the for 100 finesat different S/N ratio. Lozenges are for S/N = 5, crosses for S/N = 10. Left: an final result. This procedure is particularly unsaturated Ime ofb =21 an~ log N = 13.6; right: a saturatedfine ofb= 12 andlog N = 13.8. Small Useful to find meaningful solutions in boxes show the respecltve Ime profile.

39 .. left panel), and the fitted values are distributed around the "true" values according to Gaussian statistics (Fig. 4). For saturated lines the b - N correlation is quite strang, and the uncertainties on N are much greater. The fitted values are thus distributed along the curve of constant equivalent width, as shown in Figure 3 (right panel). The observed band 0_ .. N distributions are rather asymmetrie, with a tail extending far away from the mode value (Figs. 4a, b). In any case, most of the fitted values are still confined C) b:::~_~,.~==2c>~==::::i::~..~~ --4 -2 0 in a region close to the "true" values. oe l~g"":) Ob Different levels of S/N ratio do not seem Figure 6: Distribution ofstatisticalerrors computedbyMINUIT. Left: errors on column density; right: to change the results outlined here errors on Dopplerparameter. Solidline is from the MINOSprocedure, dashedline from the HESSE qualitatively. As an example, in Figure 5 procedure. Note that a fraction of errors estimated with HESSE are implausibly smalI. Data are for 1000 saturated lines of b = 20 and log N = 15. we report (in the N - b plane) the results of 100 simulated fits with S/N = 10 (crosses) and S/N =5 (Iozenges) for two FITLYMAN, we have performed exten­ larger and usually asymmetrie: in same different lines. At lower S/N the spread of sive tests on simulated data. Clearly, cases there is a measurable shift of the fitted values obviously increases, but these tests cannot span the whole range the mode value, even if it is always this is entirely consistent with the of resolution and signal-to-noise ratio weil within the observed 68% spread. At increased uncertainties on the equivalent available: anyway, since the test data higher column densities, however, the width. No systematic effect seems to arise were entirely generated and analysed line shape becomes unambiguously de­ as a pure effect of lower S/N. with the programmes available in the fined by the appearance of the Lorentzian Summarising, the effect of noise on the LYMAN context, we strangly invite every wings, and the fit becomes much more fitting of isolated lines is mainly a spread user to perform similar tests to evaluate precise. Finally, for the highest column of the fitted values, without any severe the possible systematic effects on his/her densities the li ne width becomes too systematic effect. For saturated, but not data. large for the Doppler braadening to be damped, lines the spread is wider, and effective, and b becomes practically un­ occurs along the curves of constant equivalent width, which is an integral Stability determined. Clearly, the values of column density and Doppler parameter for which quantity poorly affected by noise fluctua­ the lines saturate or the damping wings tions. This gives a characteristic anticor­ Concerning the stability of the algo­ appear, depend on the considered ionic relation in the band N plane. We stress rithm we found that, when single lines transition. that these simulations do not take into have been fitted, the results have praven We have investigated in more detail account other processes that may alter to be remarkably insensitive to the "first two cases, which may represent most of the fit (Rauch et al., 1993), such as guess" values used in the minimisation. the typical situations: an unsaturated selection effects in the identification of the Even implausible initial values usually Lyman-a line (b 30, log NHI 13.8) and lines, line blending and a wrang choice of converge to the correct values, the only = = a moderately saturated one (b 20, log the limits of minimisation (see below). cost being extra computing time. When = NHI 15.0). In Figure 3 the results of 1000 more complex situations - i.e. severe = simulated fits for each line are reported Error estimate blends of many lines - are to be fitted, it in the N- b plane, while in Figure 4 their may be necessary to provide reasonable band N distributions are shown. For the We have hence tested the MINUIT initial values to obtain the correct answer. unsaturated line the band N values are ability in giving realistic estimates on the This behaviour is due to the complex essentially uncorrelated (see Figure 3, statistical errors on fitted values. Errors shape of the X2 hyper-surface in the parameter space, that may present local minima for unphysical solutions. .. ..

Noise effects <> Qo<> Qo .. ::z <><> <> 12 ~ To explore the influence of noise on <> ~ Qo the fitted values we have defined a grid of ~-;: N and b values, and for each pair we <>..~ "0 <> =~oo- have computed and fitted 100 HI lines Qo<>.~ \~ with central wavelength 5000 A(Le. with .-.-+- ~: <> <> 'P"~<>_ z == 3.11) and S/N = 10. Figure 2 shows <> <>oQo -~ - the results obtained for band N (the <><> <> <> central wavelength has always been <>c:> """ <> <> <> found with great accuracy). Four regions - <> - can be identified. For unsaturated fines (Ieftmost points) the mode value is quite close to the "true" value, with a small and 1 ClQ'...... 1 ClIiil'N Figure 7: Values ofDopplerparameter band column density N obtained by the fitting procedure symmetrie scatter around it. As the on CIV doublets. Lozenges are for individually fitted lines, crosses are for the same fines whefl central f1ux of the line reaches zero, the the fit is on the two lines simultaneously. Left: b = 10 and log Nc/v= 13.8; right: b = 10 and log spread around the "true" value becomes Nc/v 14.2.

40 width, due to the reduced number of pixels: this reflects into a largerdispersion <> i i , of the fitted values. Furthermore, severe ·OO.·.....I I.!.•. 'U', . ," - ... 'I ,j systematic effects may appear, in particular in the fits of saturated lines, wh ich depend strongly on the line profile close to the continuum level. Thus, the limits must always be chosen as wide as possible, in order to follow the line profile 20 up to the continuum. Two examples of the 1_ simulated fits are shown in Figure 8.

~~1~31'--'---.~_~:l.~""'--'---.~ ~1~""...J 1 f"31...... -'--'-1--'''"-'-...... 1" :1._ 16 1 .. 1&1 lC)g...... lCllgN Improvements Figure 8: Values of Dopplerparameter band column density N obtained by the fitting procedure on Lya lines with different fit intervals. Lozenges are forlimits setathalfintensity ofthe line, crosses The Lyman context is still an evolving are for the same lines when the fit is over the whole line width. Left: b =25 and log NHI = 13.6 ; tool. Some improvements are al ready right: b =25 and log NHI = 15. For clarity, on the right only 50 representative points are plotted. being implemented, and will be released Small boxes show the respective line profile, with the different intervals considered. with the future versions of Midas. Among these, the possibility of drawing plots in the velocity space, and more commands may be computed by MINUIT either by Fit of metal systems for the research of metallic systems. The convenience of a GUI version is currently Inverting the covariance matrix (HESSE being evaluated. cOmmand) or by scanning the X2 We have also verified that the We are grateful to D. Trevese and L. hypersurface around the minimum (MI­ simultaneous fit of diff~rent metal lines Camurani for having supplied the first NOScommand) until a 1O'offset has been does significantly increase the precision version of the code, and E. Giallongo and reached.ln Figure 6 we plot the observed attainable in the fit. We have performed distributions of the errors obtained with a set of fits on simulated CIV doublet P. Petitjean for useful suggestions and discussion. both methods for 1000 simulated lines of spectra, again for different values of band b::: 20 and log N::: 15. The peak value of N, at S/N ::: 10. A self-explanatory ex­ the distributions is in good agreementwith ample for two of them is reported in Fig­ the observed dispersion in fitted values ure 7. The results of our simulation clearly References (see Fig. 4). Clearly HESSE, due to its indicate that the simultaneous fit of symmetric nature, cannot account for the different ionic transitions should be Carswell, R. F., Lanzetta, K. M., Parnell H. C., asymmetric distributions such as those exploited whenever possible. Webb J. K., 1991, ApJ 371, 36. Cristiani, S., O'Odorico, S., Fontana, A., shown in Figure 4. This is better ac­ Giallongo, E., Savaglio, S.: 1995, to be Complished by MINOS, as is shown in Limits of the fit region published in Monthly Notices of the Royal Figure 6, where the two error distributions Astronomical Society (ESO Preprint 1054). are compared: as can be seen, the peaks As has been pointed out by Rauch et al. Giallongo, E., Cristiani, S., Fontana, A, and the shapes of the MINOS errors are (1993), great care must be taken when Trevese, 0.: 1993, ApJ416, 137. 2 Mar, O.P. & Bailey, G. 1995, ProcASA, in press. not symmetric. We have also found out choosing the limits on wh ich the X is Rauch, M., Carswell, R. F., Webb J. K., that both of them are sometimes computed. Weymann, R. J. 1993 MNRAS, 260, 589. inaccurate: the first one, for inexplicable Following Rauch, we quantified the Wampler, E.J., Chugai, N.N., Petitjean, P.: reasons, provides implausibly small systematic effects by fitting the same set 1995, to be published in Astroph. Journal, values in a significant fraction (- 15 %) of of lines with different limits, repeating this April 20. (Eso Preprint 1045). the lines (see Fig. 6, dotted line), while the test for different band N values. When second is not computed in a small fraclion the limits are chosen inside the line (Le. of cases (- 5 %). Thus, we suggest the before the line has reached the con­ performance of both computations, to tinuum) two effects arise. The first is an P. Ballester check for consistency of the results. increase in the noise on the equivalent e-mail: [email protected]

Astronomers in Chile Meet at ESO in Vitacura

BQ REIPURTH, ESQ-Chile

During September 1994 the Astrono­ quently beenfairly isolated. With ourmove three-day workshop Astronomy in Chile my Support Department of ESO-Chile to Santiago this situation has significantly was held 18-20 April in Vitacura. All as­ moved its offices and facilities from the La improved, and ESO is now able to a much tronomers working in Chile were invited, ~illa observatoryto ESO's Vitacu ra prem­ larger degree to be a partner in the wider and the workshop gave the about 50 Ises in Santiago. Here a new institute of astronomical activities in Chile. attendees an opportunity to learn about astronomy has emerged and is now suc­ To celebrate ESO's new presence in the wide range of astronomical interests cessfully operating. Forthe past two dec­ Santiago and to furthertlle possibilities for currently pursued within Chile. Except the ades ESO astronomers have had their increased scientific ti es between the var­ staff working those days at La Silla, Cerro Work place at La Silla and have conse- ious astronomical institutions in Chile, a Tololo and Las Campanas, and those

41 travelling abroad, virtually all astronomers in Chile were present. The accompanying list gives an alphabeticallist of the speak­ ers and the tilles of their talks, which were held in either English or Spanish. Chilean astronomy is principally concentrated in Santiago, but recently efforts from sever­ al universities in other cities have been made to develop astronomy, and new groups are emerging in Antofagasta, La Serena and Concepci6n, led by Luis Bar­ rera, Sergio Char and Ronald Mennick­ ent, respectively. On the third day of the workshop, a special session was chaired by Mark Phillips of CTIO, who gave an introduc­ tion to the growing problem of light pollution that threatens all observatories in the world. He enumerated a number of actions that must be undertaken to safe­ guard the darkness of the night sky in Chile. During the discussion that fol­ lowed it was clear that here is a problem that will require joint efforts from all the groups affected. The workshop happened to coincide Luis Barrera (U. Cat61ica dei Norte): Fell, Felll Eduardo Hardy (U. de Chile): The Stellar and Mg 11 lines in the Spectra 01 Be Stars Populations 01 Fornax with the signing in Bonn of the supple­ Patrice Bouchet (ESO): What's new about SN Steve Heathcote(CTIO): The Herbig-Haro 47 mentary agreement between ESO and 1987A? (and Future Studies 01 Dust around Jet Chile, and Daniel Hofstadt was able to SNe) Leopoldo Inlante (U. Cat6Iica): A Survey of break the good news to the audience. The Leonardo Bronlman (U. de Chile): Molecular Faint Pairs 01 Galaxies workshop unfolded in a positive and col­ Clouds and Massive Star Formation in the William Liller (Ins!. I. Newton): Observations Galactic Disk of Novae, Dwarl Novae and False Novae laborative spirit, and we all enjoyed the Luis Campusano (U. de Chile): Large Quasar Gautier Mathys (ESO): Magnetic Field Diagno­ opportunity to share with each other our Groups sis in Ap and Bp Stars scientific results. Even the weathercollab­ EleazarCarrasco (U. Cat6Iica): CCD-Photom­ Jorge May (U. de Chile): Molecular Clouds in orated, and several warm and sunny au­ etry 01 the Brightest Galaxies in Abell the Outer Galaxy tumn days allowed the conference lunch­ Clusters Jose Maza (U. de Chile): Calan-Tololo Survey: Quasars and Seyfert Galaxies es to be held in the gardens around the Sergio Char (U. de La Serena): Studies 01 Ca II emission in Fast Rotating Stars Duilia de Mello (CTIO): Mixed Pairs of Galaxies institute. The workshop ended with uni­ Pascal Fouque (ESO): The DENIS 2 mm Sky Jorge Melnick(ESO): Star Formation in Cool­ versal agreement that more such pan­ Survey: Introduction and Present Status ing Flows Chilean astronomy meetings would be Guido Garay (U. de Chile): Recently Formed Jorge Melnick (ESO): Progress at La Silla and welcome. Massive Stars: Molecular and lonized Paranal Environs Ronald Mennickent (U. de Concepci6n): Talks Presented at the Wolfgang Gieren (U. Cat6Iica): Studies 01 Understanding Strongly Eruptive Dwarf Cepheid Variables in the Magellanic Clouds Novae Astronomy in Chile Workshop Roland Gredel (ESO): Molecular Hydrogen in Fernando Noel (U. de Chile): Sun Semidiame­ 18-20 April 1995 Herbig-Haro Objects ter Survey with a Danjon Astrolabe Adelina Gutierrez/Hugo Moreno(U. de Chile): Lars-Aake Nyman (ESO): The Kinematics 01 Thomas Augusteijn (ESO): V485 Cen: A Dwarf A Diagnostic Diagram lor Planetary Nebu­ the Bipolar Rellection Nebula IC 2220 Nova with a 59 Min Orbital Period lae and Symbiotic Stars Patricio Ortiz (U. de Chile): Search 01 Quasars using CCD's and Objective Prism Tech­ niques in Widefield Telescopes Luca Pasquini(ESO): Lithium Abundances in lhe Globular Cluster NGC 6397 Mark Phillips (CTIO): First Results from lIle High-Z Supernova Searcll Hernan Quintana (U. Cat6Iica): Cluster and Galaxy Group Mergers Bo Reipurth(ESO): Herbig-Haro Jets and Molecular Outflows Miguel Roth (Las Campanas Observatory): The Magellan Project Monica Rubio (U. de Chile): Molecular Gas in the Magellanic Clouds Maria Teresa Ruiz (U. de Chile): Cool White Dwarfs Ricardo Schmidt (CTIO): Presentlnstrumenta­ lion Projects at CTIO Robert Schommer (CTIO): The Motion 01 the Local Group with Respect to Distant Super­ novae Hugo Schwarz (ESO): M2-9: Dusty Mirrors in the Sky! Malcolm Smith (CTIO): The Gemini Project Roger Smith (CTIO): CCD's and Controllers at CTIO

42 Amateur Astronomers and Dwarf Novae

L. T JENSEN (Denmark); G. POYNER (United Kingdom); P VAN CAUTEREN (Belgium); T VANMUNSTER (Belgium)

1. Introduction the cataclysmic binaries.ln a cataclysmic superhumps, are observed in most binarya mass-Iosing secondary (often a SUUMa stars. The superhumps have a On December 15, 1855 the English late main-sequence dwarf) is in close orbit period of a few per cent longer than the astronomer John Russell Hind (1823­ with a white dwarf primary. The orbit is so orbital period. Thus the detection of these 1895) was searching for new minor close that the secondary star fills its superhumps is very important in order to planets in the constellation of Gemini. Roche-volume and hence it loses determine the orbital period of systems During this search he found a new star at material through the inner Lagrange point where this is otherwise very difficult to approximately 9th magnitude. A new L1. The lost gas is accumulated in an obtain. asteroid? He observed the starforseveral accretion disk around the white dwarf. WZ Sge type (UGWZ): This special days and found that it didn't move - thus This material then accretes onto the type of dwarf nova has outbursts very not an asteroid! The next few daysthe star surface of the white dwarf from the disko At seldom. The typical recurrence time is became fainter and fainter and it was the point where the stream of gas from the several decades. Clearly, the detection soon invisible in Hinds telescope. Hethus secondary impacts the disk, a shock front of all outbursts of these rare objects is classified the star as a faint nova. But a is formed which results in a hot spot. The extremely important. few months later in March 1856 the star luminosity of the disk and the hot spot The origin of dwarf nova outbursts is was seen again, this time by another accounts for most of the luminosity of the due to a brightening of the accretion disko English astronomer Norman Robert entire system. The orbital periods of The mass-transfer burst model proposed Pogson (1829-1891). From this moment cataclysmic binaries are very short, a little in the early 1970's by Geoffrey Bath the star was monitored more systemati­ over 1 hour to about 15 hours. As the explains the outbursts by semi-periodic cally and more outbursts were observed. orbital periods indicate, cataclysmic enhancements of the mass transfer from The period between the outbursts was binaries are very small systems, and often the secondary. calculated at approximately 100 days. A the dimensions are comparable with solar The disk-instability model proposed nova with numerous outbursts was a new diameter. Dwarf novae are cataclysmic independently by Jozef Smak and Yoji phenomenon, so the first member, U binaries exhibiting quasi-periodic erup­ Osaki in the mid-1970's is nowadays the Geminorum, of a new class of variable tions of several magnitudes (2-8). Most of most favoured model of an outburst stars, dwarf novae, was discovered! 40 the time they stay in a minimum state, but mechanism. This model explains the years passed before the next member of now and again this state is interrupted by dwarf nova outbursts as folIows: The this class, SS Cygni, was discovered in abrupt outbursts. The outbursts last from accretion disk can accumulate a certain 1896. Since then a few hundred dwarf a few days to about 14 days. The time amount of gas before it gets unstable. novae or possible dwarf novae ([Downes between the outbursts, the recurrence When instability is reached, the accretion and Shara, 1993] list 349) have been time, ranges from 10 days to months, of matter to the white dwarf increases discovered. Most of these are faint and and in some cases several years. The dramatically. We see this as an increase haveonly been observed very infrequent­ recurrence time for an individual dwarf in luminosity - an outburst. When the Iy. In several cases the classification of nova is not at all constant. It is only accretion disk has lost enough mass, it dwarf novae is based on very few possible to indicate an average recur­ becomes stable again, the increased observations. rence time, e.g. the recurrence time for accretion stops and the system returns to The observation of dwarf novae (and SS Cyg ranges from 15 to 95 days with minimum magnitude. The disk is now other cataclysmic variables) is one of the 50 being the average. The same un­ ready for a new fill-up and a new outburst few fields of astronomy in which amateur predictability holds for other dwarf nova cycle can begin. astronomers still can deliver substantial characteristics as outburst amplitude The classical novae (N) and the contributions to the work of professional and outburst duration. recurrent novae (RN) are also cataclys­ astronomers. The unpredictable behav­ Dwarf novae are divided into 4 mic variables. The outburst of classical iour of most cataclysmic variables makes subclasses according to their specific novae as weil as some recurrent novae it very difficult for professional astrono­ behaviour. are generated by thermonuclear runaway mers to monitor these variables system­ SS Cyg type (UGSS): This class reactions at the su rface of the white dwarf, atically. Our current knowledge and contains the classical dwarf novae, i.e., following long periods of accretion from understanding of the physical processes, stars like U Gem and SS Cyg. the secondary. The outburst of classical that form the basis of the dwarf novae Z Cam type (UGZ): In addition to the novae are always accompanied by an outburst mechanism, are still subject to a behaviour shown by the SS Cyg subtype, expulsion of a shell of material, whereas lot of controversy. In this article the work stars of the Z Cam type are characterised for the recurrent ones this seems not of a world-wide network of amateur dwarf by standstills in their light curves. The to be the case. A more thorough and novae observers is described. We eruption light curve is occasionally comprehensive review of dwarf novae present some examples of the important interrupted by periods of standstilllasting and other cataclysmic variables can be results achieved by these observers. from days to several years. During the found in [Patterson 1984] and the Moreover we describe the instruments, standstills the star remains at abrightness references therein. the charts and the observational methods between the normal maximum and used. minimum magnitudes. 3. The Observers, Their SU UMa type (UGSU): This subtype is Instruments and Techniques 2. Dwarf Novae characterised by frequent narrow out­ bursts, but in addition super-outbursts The major part of the visual dwarf Dwarf novae are associated with a occasionally occur. During a super­ novae observations are made by a few special type of interacting binary stars - outburst, short-period light variations, very active and enthusiastic observers.

43 EF Peg: October 1991 (Schmeer, Ger) SS UMi: August 1991 (Mitchell, UK) HV Vir: April 1992 (Schmeer, Ger) AK Cnc: January 1992 (Kato, Japan) 11 V1113 Cyg: August 1993 (Szentasko, Hungary) 12 V493 Lyr: October 1993 (Bortle, USA/Van 13 Cauteren, B) LL And: December 1 (Vanmunster, B) 14 :fY : 15 Also many stars which have been monitored by observers have been re­ 16 classified after observations had shown 17 that outburst activity wasn't quite as 18.1.- ...... ---'- ---' "- ...... predicted. If a programme star is found to go in outburstfrequently, it isdropped from 2449580 2449585 2449590 2449595 2449600 2449605 2449610 the programme, but only after monitoring JD it intensely through several outbursts. Figure 1. This has recently happened with the UGSU star SS UMi. Most of them are very experienced observers contribute to the dwarf nova Figure 1 shows an example of a light variable star observers, with proven skills section of the AAVSO Circular. Often an curve obtained for a programme star: in 'traditional' variable star work. Lot of outburst is only observed by a few UGWZ type dwarf nova UZ Bootis. The practice and perseverance has turned observers. Nearly 400 variables are light curve documents arecent superout­ them into well-trained amateurs, that are currently classified as UG stars. For many burst (August 1994). The last observed capable of memorising tens of star fields. possible UG stars nearly no characteris­ outburst of UZ Boo was in 1978. On every clear night, they visually check tics are known, because they are very Occasionally stars which are not erup­ these fields, hunting for 'new' stars. No faint even in outburst or they have very tive in nature, but were originally thought sophisticated instruments are required: long recurrence times. Some of them to be are revealed. Four of these rogue an alt-azimuth dobsonian type of reflector have only been observed on very few stars have recently been identified and is quite common, with apertures in the occasions, for some the identification is dropped from the programme. Three, HN range from 20 cm to 50 cm and more. unknown, or the subclass classification is Cyg, UYVul and UZVul, were infactsemi­ Using this technique, some amateurs are very tentative. The task of monitoring regular type variables catalogued as capable of visually inspecting between 50 such stars is largely taken over by possible cataclysmic stars. However, ob­ and 100 star fields in one single night! On amateurs, who join their efforts in world­ servations from the programme's observ­ such nights typically 3-10 dwarf novae wide networks, that in total coverover 100 ers have revealed their real nature. When are seen in outburst. peculiar objects. All over the world, only a activity of a suspected star is noticed, Often the magnitude of a dwarf nova in few dozens of amateurs participate in other members of the network are in­ minimum is in the range magnitude 17­ these dwarf novae alert networks. formed (overtelephone) in order to obtain 21. Most dwarf novae are hence only immediate confirmation. Positive identifi­ observed in maximum and naturally the cations of outbursts result in alert calls observers get a lot of negative (fainter 3. Interest from Professional (telegrams and electronic circulars), that than) observations. Negative observa­ Astronomers are issued to amateurs all over the world tions are often as important as the positive and to interested professional astrono­ ones to get a continuous Iightcurve. What The Astronomer organisation (UK) mers. makes dwarf novae hunting so challeng­ has set up a programme specifically Examples of such services are The ing to amateur astronomers? There is no to monitor poorly studied-Iong period Astronomer Electronic Circulars (UK) and single answer to this question. Most dwarf novae. The Recurrent Objects the Cataclysmic Variables Circular (B). amateurs are driven by the unpredictable Programme contains approximately 75 Photometric observations, that cover the nature of the dwarf novae: no one can objects of various types, ranging from outburst activity, are passed between anticipate the outcome of a nightly dwarf novae, recurrent and old novae, to professional and amateurastronomers by observation session. Others are attracted suspected UG, UGSU and UGWZ stars. electronic mail, mainly through Internet by the 'scientific' contribution of their Gary Poyner has been coordinator of the and CompuServe. The availability of e­ observations. Variable star charts and programme since 1991. The main criteria mail to many amateurs has meant that sequences come from various organisa­ in the object selection were the following: professional astronomers can be alerted tions, such as the AAVSO (American 1. The star must have an outburst to an outburst very quickly, thus making Association of Variable Star Observers), period (or suspected period) of at least possible extremelyval uableobservations The Astronomer and the BAAVSS (UK), one year. early in the outburst phase and in various RASNZ (Royal Astronomical Society of 2. Little (if any) information on the wavelengths.This can provide valuable New Zealand), BVVS (Belgium Vereni­ precise cycle length or amplitude is information into the mechanisms behind ging voor Sterrenkunde), and from available. these cataclysmic events. Summarising reprints of articles in professional journals Gradually more stars were added to results, complemented with contributions [Vogt and Bateson, 1982], [Bruch et al. , the list, as research uncovered many from professional astronomers, are pub­ 1987] and [Downes and Shara, 1993]. more of these long-period objects where lished in various magazines. Monthly tables, listing all dwarf novae just one or perhaps two outbursts had outburst activities are generated and been recorded. Among recent successes 4. 1. HV Virginis distributed by organisations like the we note the first ever visual observations AAVSO (The AAVSO Circular). Approxi­ of the following stars (discoverers' names The HV Vir outburst deserves special mately 40-50 more or less active following). mention here, as it proves just how

44 valuable amateur observations can be. programme, complementary to the TA observer, before the astronomical world is HV Vir was photographed by Schneller in programme. It mainly consists of poorly informed. Confirmation via a CCD image 1929 in outburst at magnitude 11.5. observed objects, for wh ich professional is much more 'safe' than a second visual Following aseries of observations of it at astronomers have shown interest and sighting. Having something on file/paper maximum light and in decline, it was have requested continuous monitoring. also makes it possible to analyse and re­ classified as a classical nova. The object The CVAP currently includes about 13 check the result later on. was added to the recurrent objects objects, of wh ich the identification and/or programme in 1988 primarily because of the subclass is uncertain or unknown. (3) Astrometry/photometric sequence it's high galactic latitude, and the possibil­ These objects are now monitored on a ity that it may be a rare type WZ Sge dwarf very regular basis by European variable With CGO images an accurate position nova, although no other outbursts had star amateur observers. Oue to the measurement may be obtained. At been detected. On April 20, 1992, a unavailability of professional search present, too many inaccurate charts are German amateur, Patrick Schmeer, made charts and sequences for these objects, used, resulting in 'false' outburst-alerts or his usual check of the field with his 20-cm the CVAP members currently are using large scatter in magnitude estimates. telescope and saw HV Vir in outburst at charts based on the Guide Star Catalogue Preparing charts with reliable comparison magnitude 12. At first Schmeer thought it (GSC). Example CVAP objects include: stars is of utmost importance. might be a minor planet, but following a AS Psc (last seen in outburst in 1963), search of minor planet positions it soon V358 Lyr (seen in 1965) and SS LMi (seen (4) Photometry became evident that HV Vir was indeed in in 1980, and classified as either N or UG). By taking regular images of the Although the CVAP started only re­ outburst. Observations of spectra and variable star, reliable photometry may be high-speed photometry have shown that cently, a first result has al ready been performed. The accuracy is much higher HV Vir is indeed a dwarf nova. It shares obtained. The programme star IR Lyr, than the visual estimates. Oetecting c1assified as a possible UG star, has manycharacteristics ofWZ Sge itself, and superhumps in suspected UGSU stars is been observed during activity in recent displays superhumps typical of UGSU an example of useful work that can be months. A first interpretation of photo­ stars. done. metric observations indicates that the It is our impression that most astron­ star does probably not belong to the UG 4.2. LL Andromedae omers do not know about the important variables. Paul Van Cauteren and Tonny work some amateurs do in the field of Vanmunster are currently the coordina­ Another major success of the TA dwarf cataclysmic variables. It is our hope that tors of the Belgian programme. nova network was the detection of an this article will inspire more professional outburst of LL And by Belgian amateur astronomers to make use of the results Tonny Vanmunster on the evening of and observations obtained by the ama­ 6. Future Perspectives Oecember 7,1993 using a 35-cm dobso­ teur astronomers. The different recurrent nian reflector. The outburst was con­ object programmes are always open for As CCO cameras become more widely firmed within minutes by UK and Belgian new interesting cataclysmic variables to available to amateur astronomers, we observers. LLAnd was discovered photo­ be monitored. Naturally, it is also in­ would like to point out the important role graphically by astronomer Paul Wild spiring for the amateurs to see that that these instruments can play in the (Bern Observatory) in September 1979 their efforts can be and are used by observation of dwarf novae, even if they [Wild 1979], when it reached V-13th. professional researchers to gain more are mounted on a small telescope. Since then, the dwarf nova had never knowledge about the behaviour of cata­ been seen again. Oue to an alert, the clysmic binaries. Oecember 1993 (super)outburst could (1) Monitoring of dwarf novae also be monitored by several profession­ at quiescence al astronomers. Spectroscopic observa­ References tions of the star at maximum light were Most amateur astronomers, even with Patterson 1984: J. Patterson, The evolution 01 obtained, as weil as a photometric larger instruments, obtain only limiting cataclysmic and low-mass X-ray binaries. outburst light curve based mainly on magnitudes around magnitude 14.5­ The Astrophysicat Journat Suppt. Ser.. 54, observations contributed by amateurs. 15.5. But the majority of dwarf novae are p. 443-493, (1984). The outburst amplitude was - 6 mag­ much fainter when at quiescence. With Vogt 1982: N. Vogt and F.M. Bateson, An atlas 01 southern and equatorial dwarf novae, nitudes. Or. Kato (Kyoto University, Ja­ the use of a CCO camera, however, many Astron. Astrophys. Suppt. Ser., 48, p. 383­ pan) conducted CCD photometry during of these objects are within reach. This 407, (1982). four nights, and detected superhumps, means that daily monitoring may result in Bruch 1987: A. Bruch, F.-J. Fischer and U. hence classifying the star as a UGSU observing the start of the outbursts, and Wilmsen, An atlas and catalogue 01 north­ dwarf nova, possibly belonging to the WZ early warning of interested observers is ern dwarf novae, Astron. Astrophys. Suppt. Ser., 70, p. 481-516, (1987). Sge subclass. possible. Downes 1993: RA Downes and M.M. Shara, A Catalog and Atlas 01 cataclysmic vari­ 5. Belgian Cataclysmic Variables (2) Gonfirmation of outburst ables, P.A.S.P., 105, p. 127-245 (1993). Alert Programme discovered by visual observers Wild 1979: lAU Circular No. 3412.

The Belgian Cataclysmic Variables When an outburst is reported by a E-mail address: Alert Programme (CVAP) was initiated in visual observer, it is common practice to Lasse Teist Jensen: May 1994 as a dwarf nova observation obtain confirmation from a second [email protected]

45 ANNOUNCEMENTS

08.06.1995 Telescope Software Scientist (CTR 118) Grade 8 La Silla Observatory, Chile. This post is open to suitably qualilied men and women. Education: University degree or equivalent in computer science or related lield. Experience and knowledge: Experience in UNIX systems and soltware. A knowledge 01 C and Fortran is essential, and the candidate should have some experience with workstations, PCs and Macintosh computers. Some knowledge 01 VME or equivalent data bus systems would be an advantage. Fluency in English is required. Assignment: The successlul candidate will be part 01 the team 01 ten scientists and engineers responsible lor the day-to-day operation and continuing development 01 the Swedish-ESO (Sub) Millimetre Telescope (SEST) and its operating system atthe La Silla Observatory in Chile. The schedule 01 work requires that he/she be capable 01 working both independently and in a group. Duty station: La Silla, Chile. Starting date: As soon as possible. Remuneration: The remuneration lorthispost will becommensuratewith the background, experienceand lamilystatus. The basic monthly salary (tax-Iree) will not be less than DM 6.072,-. Furthermore, an expatriation allowance 01 either 30-35% (single) or40-45% (head ollamily) 01 the basic salary as weil as some other allowances may be added. Applications should be submilted before July 20, 1995. Application forms may be obtained Irom ESO, Personnel Services, Karl-Schwarzscllild-Str. 2, 0-85748 Garching bei München, Germany.

ESO Fellowships in Chile 1996/97 The European Southern Observatory (ESO) intends to award one or more postdoctoral lellowships tenable at ESO's Astronomy Centre in Santiago, Chile. The main areas 01 activity 01 ESO in Chile are: • Research in observational and theoretical astrophysics; • Operating the La Silla Observatory; • Building the VLT, consisting 01 lour 8-m telescopes, at Paranal. Fellows will spend most 01 their time at ESO's centre in Santiago, but will regularly visilthe La Silla Observatory, where they will take part in supporting visiting astronomers and maintaining the instrumentation. Fellows will normally spend up to 50% 01 1I1eir time in support work, and the remainder doing their personal research. The lellowsllip programme in Chile olfers a unique opportunity to learn and to participate in lhe process 01 observational astronomy while pursuing a research programme wilh state-ol-the-art lacilities. The La Silla Observalory has eight telescopes in the range 0.9 m to 3.6 m, as weil as the 1-m Schmidt, the 15-m SEST millimetre radio telescope and smaller instruments. ESO's Institute 01 Astronomy in Santiago oHers computational lacilities, a library, as weil as the stimulation 01 colleagues Irom ESO and the local Chilean astronomy community. Applicants should have arecent doctorate. The basic monthly salary will be nol less than DM 5526 to which is added an expatriation allowance 01 up to 40%. The lellowships are granted lor one year with the expectation 01 a renewal lor a second year and exceptionally a third year. Fellowships begin between April and October 01 the year in which they are awarded. Applications should be submilted to ESO not later than October 15, 1995. Applicants will be notilied in Oecember 1995. The ESO Fellowship Application Form should be used. The applicanl should arrange lor three lelters 01 recommendation lrom persons lamiliar with the scientilic work 01 the applicant to be sent direclly to ESO. These lelters should reach ESO not later than October 15,1995. Inquiries, requests lor application lorms, and completed applications should be addressed to: European Soulhern Observatory, Fellowship Programme, Karl-Schwarzschild-Str. 2, D-85748 Garching bei München, Germany

ESO Fellowships in Garehing 1996/97 The European Southern Observatory (ESO) intends to award up to six post-docloral lellowships tenable atthe ESO Headquarters, located in Garching near Munich. The main areas 01 activity at the Garching Headquarters are: • Research in observational and theoretical astrophysics; • Managing and building the Very Large Telescope (VLT); • Oevelopment 01 instruments lor current ESO telescopes and lor the VLT; • Management, calibration, analysis and archival 01 astronomical data lor the current ESO telescopes and the VLT; • Fostering co-operation in astronomy and astrophysics within Europe. Fellows normally participate in one or more 01 the above activities. There is also the possibility 01 participating in the activities 01 the Space Telescope European Coordinating Facility (ST-ECF), which is located in the ESO Headquarters building. In addition to personal research, Fellows will spend up to 25 % 01 their time in support or development activities related to the main areas 01 activity atthe ESO Headquarters. This functional work will normally include a lew short-term visits per year to Chile (La Silla and Vitacura). ESO lacililies include the La Silla Observatory in Chile with its eight telescopes in the range 0.9 m to 3.6 m, as weil as the 15-m SEST millimetre radio telescope, and smaller instruments. In Garching, extensive measuring, image processing and computing lacilities are available. There are several Max-Planck Institutes and the University Observatory in the Munich area with major programmes in astronomy and astrophysics and with which joint programmes can be conducted. Applicants should have arecent doctorate. The basic monthly salary will be notless than DM 5526 to which is added an expatriation allowance 01 9-12% il applicable. The lellowships are granted lor one year with the expectation 01 a renewal lor a second year. Applications should be submilted to ESO not later than October 15, 1995. Fellowships begin between April and October 01 the year in which they are awarded. Applicants will be notilied in Oecember 1995 or soon thereatter. The ESO FellowshipApplication lorm should be used. The applicant should arrange lor three lelters 01 recommendation Irom persons lamiliar with the scientilic work 01 the applicant, to be sent direclly to ESO. These lelters should reach ESO not later than October 15, 1995. Inquiries, requests lor application forms, and completed applications should be addressed to: European Southern Observatory, Fellowship Programme, Karl-Schwarzschild-Str. 2, 0-85748 Garching bei München, Germany

46 ------Postdoctoral Fellowship on La Silla - NTT Upgrade Project In the framework ofthe Very Large Telescope Programme, ESO has undertaken a projectto upgrade the 3.5-m NewTechnologyTelescope (NTT) in order to test operationai concepts and software for the VLT. As part of this development, ESO is oftering a fellowship to a qualified opticai astronomer who would like to participate in this programme. In addition to carrying out an independent research programme (50%), specific duties (50%) would include support of Visiting Astronomers and calibration and performance control of the instrumentation on the NTT. This position is intended to ofter the recipient the opportunities both to develop an independent research programme with the facilities of a major observatory and to contribute to the realisation of the NTT Upgrade Project. The successful candidate will be expected to work in close collaboration with the scientists and engineers of the NTT Team to ensure the success of the NTT upgrade project. Scientifically, collaboration with the Astronomy Support Department (ASO) of ESO-Chile is encouraged. Current research interests within the ASO are: active galactic nuclei, star formation, supernovae, RR Lyrae stars, chemical abundances, the interstellar medium, the activity of cool, and magnetic stars. Knowledge of modern software utilities is a requirement. Candidates acquainted with (surface-) photometric techniques are especially encouraged to apply. The ESO fellowships are granted for aperiod of one year, renewable for a second year and exceptionally for a third year. The monthly basic salary will not be less than DM 5279, to which are added an expatriation allowance of 30-45% as weil as a mountain allowance of 5-10%. Starting date: As soon as possible. Applications should be submitted to ESO not later than July 20,1995. Applicants will be notified by October 31, 1995. Application forms are available from ESO Personnel and General Services (PGS), Karl-Schwarzschild-Str. 2, 0-85748 Garching bei München, Germany. Applicants should arrange for 3 letters of reference to be sent by the same date directly to PGS. For further information: contact the NTT upgrade project scientist (Internet: [email protected]). ---- -~-~--~------~---~------

------~-~------Staff Astronomer (ESO 208) 08.06.1995 Science Division at the ESO Headquarters in Garching near Munich, Germany. Grade 9/10 This post is open to suitably qualified men and women. Education: PhD in Astronomy, Astrophysics or Physics. Experience and knowledge: Candidates must have several years of postdoctoral research experience, and have contributed significantly to at least one area of modern astrophysics. They must also have substantial experience in the use of large ground-based telescopes. Evidence of an interest and ademonstrated ski 11 in exploiting telescopes and their instruments to their limits will be looked for. Assignment: The successful candidate will be expected to carry out a significant programme of personal research for up to 50% of the time. The appointee will also work in the Science Division with the Associate Director for Science (J. Bergeron) in strengthening interactions between the Science Division and the VLT and Instrumentation Divisions. The appointee will provide scientific input for VLT instrumentation in particular for pipe-line calibrations and archiving of the optical spectrographs FORS and FUEGOS. The appointee will also supervise the VLT instrumentation-related functional work carried out by ESO Fellows. Three letters of recommendation from persons familiar with the scientific work and observational experience of the applicant should be sent to ESO, Personnel Services, directly. Duty station: Garching near Munich, Germany. Starting date: As soon as possible. Contract: This position is a three-year, renewable contract and may lead to a tenure staft appointment. Serious consideration will be given to outstanding candidates willing to be seconded at ESO on extended leaves from their home institutions. Remuneration: The remuneration forthis postwill be commensuratewith the background, experience andfamilystatus. The basic monthly salary (tax-free) will be in the range of DM 6.845,- to DM 9.583,-. Furthermore, an expatriation allowance as weil as some other allowances may be added. Application forms may be obtained from ESO, Personnel Services, Karl-Schwarzschild-Str. 2,0-85748 Garching bei München, Germany. Applications should be submitted before 20 July 1995.

1077. PA Shaver: High Redshift Quasars. Invited paper presented at New ESO Publications the 17th Texas Symposium, 12-16 Dec. 1994; to appear in 17th Texas Symposium on Relativistic Astrophysics and Cosmology (March - May 1995) (ed. H. Böhringer et al., Ann. New York Academy of Science). 1078. N.Y. Lu and W. Freudling: Large-Scale Structures in the Zone of Scientific Report No. 16: Fourth Catalogue of Stars Avoidance: The Galactic Anticenter Region. ApJ. Measured in the Long-Term Photometry of Variables Project (1992­ 1079. G. Carraro and F. Patat: The Stellar Content of the Open Clusters 1994). Tombaugh 1 and Rupprecht 46. M.N.RAS. 1080. P. Frisch et al.: Evolution of the Supercluster-Void Network. AA. Scientific Preprints 1081. M. Bobrowsky et al.: He 3-1475 and its Jets. 1082. H.E. Schwarz, L.-A. Nyman, E.R. Seaquist, R.J.lvison:ASearch 1068. F. Murtagh, J.-L. Starck, A. Bijaoui: Image Restauration with for SiO Maser Emission from Symbiotic Miras. AA. Noise Suppression Using a Multi-Resolution Support. AA. 1083. W. Freudling et al.: Determination of Malmquist Bias and 1069. U. Lindneretal.: The Structure of Supervoids-I. Void Hierarchy Selection Eftects from Monte-Carlo Simulations. A.J. in the Northern Local Supervoid. AA. 1084. G. Meylan, M. Mayor, A. Duquennoy, P. Dubath: Central Vlocity 1070. S. Cristiani et al.: The ESO Key-Programme "A Homogeneous DisperSion In the Globular Cluster Ol Centauri. AA. Bright QSO Survey" - I. AA. . 1085. F. Courbin et al.: Photometric Monitoring (1987 to 1994) of the 1071. J. Rönnback and N. Bergvall: Blue Low Surface-Bnghtness Gravitational Lens Candidate UM 425. AA. Galaxies. 11. Spectroscopy and Chemical Abundances. AA. 1072. J.K. Kotilainen et al.: The Nature of the two Nuclei in the Young Technical Preprints Merger NGC 3256: An Obscured AGN? AA.. . 1073. D. Minniti: Spectroscopy and IR Photometry for Glant Stars In 66. S. D'Odorico: Array Detectors and Instruments for the ESO Obscured Globular Clusters: NGC 6325, NGC 6401, NGC VLT. 6440, NGC 6517, NGC 6642, HP1 and PAL6. AA. Highlights of lAU Symposium No. 167. 1074. D. Minniti: Abundances and Velocities for Open and Globular Review papers to be published in the Proc. of the lAU Symp. Giants: The Data. AA. No. 167, "New Developments in Array Technology and 1075. N.N. Chugai, I.J. Danziger, M. Della Valle: Optical Spectrum of Applications". SN 1978K: Emission from Shocked Clouds in the Circumstellar 67. Very Large Telescope - Instrumentation. Wind. M.N.RAS. Papers submitted to the International Conference "Scientific 1076. F. Murtagh, A. Aussem, M. Sarazin: Nowcasting Astronomical and Engineering Frontiers for 8-1 Om Telescopes". 4-60cto­ Seeing: Towards an Operating Approach. PA.S.P ber1994, Tokyo,Japan.

47 STAFF MOVEMENTS Departures Europe Arrivals LAMADIE, Micheie (F), Secretary PELLEGRINI, Silvia (I), Fellow Europe RASMUSSEN, Bo Frese (OK), Tech- COTE, Stephanie (CON), Fellow nician (Software) EGEOAL, Carsten (OK), Engineer THIMM, Guido (0), Fellow SACRE, Philippe (B), Mechanical WAMPLER, Joseph (USA), Scientist Engineer STRASSER, Jose! (0), Contract Officer Chile VAN DER STROM, Margaretha (NL), GIRAUO, Edmond (F), Associate Contract Officer VAN DER BLlEK, Nicole (NL), Student

Contents

R. Giacconi: Recent Developments .

TELESCOPES AND INSTRUMENTATION

M. Tarenghi: News from the VLT Programme 2 J. Andersen: Scientific Priorities for La Silla in the VLT Era 4 M. Faucherre: Is the Seeing Situation at the 3.6-m Telescope Irreversible? 5 L. Pasquini, L. Kaper: CAT/CES NEWS 9 G. Rupprecht: The FORS Focal Reducers for the VLT - a Status Report 9 H. Dekker: UVES (UV-Visual Echelle Spectrograph) tor the VLT - a Status Report...... 11 P.O. Lagage, Y. Rio, D. Dubreuil, Th. De Graauw, J.w. Pel, 1. Shoemaker: Re- sult of the Phase A Study for the VLT Mid-Infrared Instrument: VISIR 13 ND Bits & Pixels 16 VLT Insurance Contract Signed 18

SCIENCE WITH THE VLT

A. Maeder, G. Meynet: Topical Astrophysical Problems on Massive Stars for VLT Observations 19

REPORTS FROM OBSERVERS

D. Alloin, M. Santos-L1e6, G. Stirpe, B.M. Peterson: Monitoring of Active Galaclic Nuclei: the Why and the How...... 25 M.E. Giannuzzo, G.M. Stirpe: On the Variability of Narrow-Line Seyfert 1 Galaxies 28 C. Alard, J. Guibert, O. Bienayme, D. Valls-Gabaud, A.C. Robin, A. Terzan, E. Bertin: The DUO Programme: First Results of a Microlensing Investiga­ tion of the Galactic Disk and Bulge Conducted with the ESO Schmidt Telescope 31 G. Burki, F. Rufener, M. Burnet, C. Richard, A. Blecha, P. Bratschi: The Varia- tion of Atmospheric Extinction at La Silla 34

OTHER ASTRONOMICAL NEWS

A. Fontana, P. Ballester: FITLYMAN: AMidas Tool for the Analysis of Absorp- tion Spectra 37 Bo Reipurth: Astronomers in Chile Meet at ESO in Vitacura 41 L.T. Jensen, G. Poyner, P. van Cauteren, 1. Vanmunster: Amateur Astronom- ers and Dwarf Novae 43

ANNOUNCEMENTS

Telescope Software Scientist (CTR 118)...... 46 ESO Fellowships in Chile 1996/97 46 ESO Fellowships in Garching 1996/97 46 Postdoctoral Fellowship on La Silla - NTT Upgrade Project 47 Staft Astronomer (ESD 208) 47 New ESO Publications 47 Staff Movements 48

48