The Messenger

No. 142 – December 2010 Improving network links to Paranal to links network Improving Facility Optics Adaptive The VLT populations in globular clustersMultiple 1 Westerlund cluster super Dissecting the The Organisation

Enabling Virtual Access to Latin-American Southern Observatories

Giorgio Filippi1 The partners of EVALSO are the Università The communications technology degli Studi di Trieste, Italy; the European Southern Observatory; Ruhr-Universität To guarantee current needs while making 1 ESO Bochum (RUB), Germany; Consortium provision for future performance demands, Gestione Ampliamento Rete Ricerca the EVALSO project has opted for optical (GARR), Italy; Universiteit Leiden, the fibre network paths that could be procured EVALSO (Enabling Virtual Access to ­Netherlands; Istituto Nazionale di Astro­ from commercial installations or, where Latin-American Southern Observatories) fisica (INAF), Italy; Queen Mary, University none existed, created new ones. The use is an international consortium of nine of London, United Kingdom; Cooperación of dark fibres (i.e. physical fibres whose full astronomical organisations and research Latino-Americana de Redes Avanzadas capacity is available) and, where this is not network operators, part-funded under (CLARA), Uruguay; and Red Universitaria economically possible, use of reserved the European Commission FP7, to create Nacional (REUNA), Chile. More details wavelengths, provides an optimal and flex­ and exploit high-speed bandwidth con- of the overall project and its members are ible infrastructure for the traffic manage­ nections to South American observato- available at the EVALSO website1. ment needs of today, with a clear potential ries. A brief description of the project for future growth. In order to exploit these is presented. The EVALSO Consortium The project focuses on two aspects: optical communications most effectively, inaugurated a fibre link between the – building an infrastructure to connect DWDM (dense wavelength division multi­ ­Paranal Observatory and international observatory sites efficiently to the plexing) technology was selected for the networks on 4 November 2010 capable European astronomical community by communications equipment. of 10 Gigabit per second. linking to the network infrastructures ­created in recent years with EC support The optical path infrastructure (numbers (in particular ALICE 2 and GEANT 3); refer to Figure 1) consists of: It should come as no surprise that the – promoting research activities to enable – new fibre cables (1 and 4) that serve the remoteness which makes a site excellent and validate new ways to interact Paranal Observatory and OCA sites up for optical–near-infrared astronomy also with the remote facilities made available to Ruta 5 at La Varilla; conflicts with the need for good con­ by high-bandwidth communications. – a pair of fibres from the existing installa­ nections to a fast communications infra­ More details can be found in Filippi et al. tion along Ruta 5 to relay from La Varilla structure. With the constant increase (2010). to Antofagasta; in data rates and, more generally, in com­ – fixed wavelength communication (2) munications needs, limits in bandwidth The present article focuses on the com­ between the telecommunications pro­ may impede the efficiency of operations, munications infrastructure, which is vider Point of Presence (TELCO PoP) in and hinder future expansion. For the broken down into two major components: Antofagasta and Santiago; ESO Paranal Observatory and the Ruhr- the optical paths and the communica- – dark fibres (5) between the TELCO PoP Universität Bochum Observatorio de Cerro tions equipment. The locations served by and the end points in Santiago, namely Armazones (OCA), this conflict will soon the new high capacity infrastructure are: ESO’s Vitacura offices and the REUNA cease to exist. On 4 November 2010, at the ESO Paranal Observatory and OCA, office in Providencia; the ESO Vitacura offices in Santiago, the about 1200 km north of Santiago; the – and housing space for EVALSO equip­ EVALSO Consortium formally inaugurated REUNA offices in Antofagasta (about ment at the TELCO PoP in Antofagasta a fibre-based system capable of 10 Giga­ 120 km from Paranal); and the ESO and (6) and Santiago (3). bit per second (Gbps) to connect the two REUNA offices in Santiago de Chile. observatories to the international aca­ In order to connect the two observatories demic networks. to the Chilean TELCO infrastructure, new cables had to be installed in the Ata­ Figure 1. Overview of the EVALSO communications cama Desert for a total of about 100 km. infrastructure. See text for details. The cables were of the type used for direct underground installation and the first cable, approximately 80 km long, was 4 laid down between Paranal and the point of connection to the telecommunications provider’s backbone, situated along the 6 Atacama 1 7 Panamericana (Ruta 5, the Pan-American Highway) at La Varilla. The second, ap­ 2 proximately 20 km long, is between OCA and an intermediate point on the first 3 cable. These installations established an optical fibre path between the two obser­ 5 vatories and, together with the portion Dark fibre: unlimited capacity of fibre procured from the commercial net­ Lambda: (currently) 10 Gbps work, to Antofagasta.

2 The Messenger 142 – December 2010 Underground installation is not only more traffic is passed to the provider equipment illustrated by Giorgio Filippi, ESO, for the secure, but limits the visual impact on for the approximately 1200 km section optical paths, and Sandra Jaque, REUNA, the pristine desert environment. A special over the LAMBDA cable to Santiago. The for the DWDM equipment. machine cuts a trench about 20 cm wide LAMBDA termination in Santiago is the and 80 to 120 cm deep, depending on the distribution point for the traffic coming The overall system is now available for nature of the soil (see Figure 2). The cable from the northern region of Chile. Using the research activities of the EVALSO is suitable for direct installation without local dark fibres, the Paranal traffic is Consortium. There are three key areas that ducts (Figure 2). As reels of 4 km length ­delivered to the ESO Vitacura offices and the project will develop: are used, splicing boxes have to be placed the REUNA traffic to the REUNA offices – Fast and efficient access to data: at pre­ in chambers at the same distance. The in Santiago, and from there to the com­ sent, data collected at the Paranal Ob- construction also crosses existing roads mercial and academic networks. servatory can take anywhere from hours and a gas pipeline. A movie of the cable- to weeks before it becomes available for laying process is included in the ESO pod­ use in Europe. EVALSO will allow data to cast4. The distance from the end of the Project history and inauguration be transferred in near real time. new cable, at the La Varilla crossroads, – Virtual presence at the observatories: to the nearest installation of the telecom­ The EVALSO project team (SA1 work the bandwidth offered by EVALSO munications provider in Antofagasta is package) began work in 2008 by taking makes a “virtual presence” at the obser­ about 50 km. This distance has been cov­ over an initial market survey, which gave vatories possible. ered by procuring a fibre pair from the enough confidence in terms of technical – New possibilities in observing: EVALSO existing facility. A fixed wavelength channel and economic feasibility to continue with can serve several of the needs of its (LAMBDA), capable of carrying 10 Gbps, the project. Based on this input, the SA1 partner members, for example, by oper­ links Antofagasta and Santiago. team detailed the specification of the final ating robotic telescopes on Cerro system and entered the procurement ­Armazones, supporting their use for EVALSO equipment is installed at five phase, assigning the optical paths to ESO educational purposes and improving the nodes. At Paranal this node receives the (in coordination with OCA for their part scientific exploitation of the capabilities traffic to and from both observatories. of the connectivity) and the DWDM equip­ of both observatories through innova- For each user community separate sub- ment to REUNA. The complex procure­ tive operations schemes, which will act channels are used to transport the traffic ment phase took nearly the whole of 2009, as pathfinders for the operation of future to its final destination. The aggregation leading to the final decision and the start facilities like the European Extremely point for the traffic handled by REUNA and of engineering in the last quarter. The Large Telescope. from the observatories is sited at Antofa­ installation of the new cables as well as As a member of the Consortium, ESO gasta. Here the bundled (but not mixed!) the configuration and integration of the is planning to integrate the ­system into its existing commercial parts, both paths and communications infrastructure in the DWDM equipment, kept the team busy ­coming months, leading to its use in regu­ until October 2010, when the first trans­ lar operation of the telescopes on Paranal. mission test could be made. Results of these developments will be The EVALSO formal inauguration took reported in a future article. place in a ceremony at the ESO Offices in Santiago (see the picture on the Astro­ nomical News section page 40). The event References was attended by a number of ambassa­ Filippi, G. et al. 2010, Proc. SPIE, 7740, 77401G dors and diplomats from EU countries, a very large delegation from the EC, high level officials from ESO, ALMA and other Links organisations. The event was opened by a 1 EVALSO web page: http://www.evalso.eu short speech by the ESO Director General, 2 ALICE network: http://www.redclara.net Tim de Zeeuw, and there were speeches 3 GEANT pan-European network: http://www.geant. from Josè Palacios (REUNA president), Rolf net/About_GEANT/pages/home.aspx 4 Chini (RUB-OCA), Massimo Tarenghi (ESO ESO Podcast: http://www.eso.org/public/videos/ eso1043a/ representative in Chile), Mario Campolargo (EC) and Ambassador Fernando Schmidt Ariztía (vice-minister of Foreign Affairs of Chile). Fernando Liello, Università degli Studi di Trieste, EVALSO Project Coordi­ nator, explained the EVALSO framework Figure 2. Two views of the work involved in laying the and its relations with the overall aca­- EVALSO optical cable from Paranal to link with the existing network infrastructure; upper image: cutting demic research networks. The technical the trench; lower: laying the cable in the trench. aspects of the EVALSO infrastructure were

The Messenger 142 – December 2010 3 Telescopes and Instrumentation (p. 12) for details. al. et Arsenault See testing. in shown mirror, 1.1-metre lightweighted aZerodur mirror, ondary The Adaptive Optics Facility deformable sec plateau. See Testi et al. (p. details. (p. al. more 17) et Testi for See ­plateau. Chajnantor the on site high the at ferometer inter an as observing antennas ALMA Eight ­ ­

Credit: ALMA (ESO/NAOJ/NRAO) Telescopes and Instrumentation

On the Instrumental Polarisation of NAOS–CONICA

Gunther Witzel1 module NAOS and the camera CONICA. A model of the instrumental polarisation Andreas Eckart1 This instrument provides a mode for Rainer Lenzen 2 polarimetric differential imaging combin­ To understand the IP of NACO we have Christian Straubmeier1 ing a Wollaston prism and a l/2 wave developed a model of the instrumental plate (a half-wave plate, HWP). The Wol­ effects within the Stokes/Mueller formal­ laston prism separates the incoming par­ ism. The polarisation of the incoming light 1 I. Physikalisches Institut, Universität tially polarised light into two orthogonal, is mainly affected by reflections at metal­ Köln, Germany linearly polarised outgoing beams (the lic coatings of mirrors within the light path 2 Max-Planck-Institut für Astronomie, ordinary and extra-ordinary beams). The of NACO. Each reflection causes cross­ ­Heidelberg, Germany HWP enables the observer to change talk between the four Stokes parameters the angle at which the measurement is that describe the polarisation state. For conducted without rotating the Wollaston unpolarised incoming light this crosstalk As expected for a Nasmyth instrument, prism with respect to the detector and generates, for example, linear polarisa­- NAOS–CONICA shows a significant thus without changing the field of view. tion orthogonal to the plane of incidence. instrumental polarisation that is strongly Intensity measurements over at least four The magnitude of the crosstalk depends dependent on the parallactic angle angles (0°, 45°, 90° and 135°) in two inte­ strongly on the angle of incidence: only of the source and can reach up to 4 % grations (two orthogonal beams simulta­ mirrors with an inclination of > 20° will in the degree of linear polarisation. neously) are necessary to obtain infor­ contribute significantly to the IP. All polari­ Detailed modelling of the polarimetric mation about linear polarisation. Circular metric effects of the optical elements and properties of the optical components polarisation cannot be measured, since their relative orientation can be described of NAOS–CONICA using the Stokes/ the polarimetric mode of NACO does not by linear operations on the Stokes vector, Mueller formalism allows the instrumen- include a l/4 wave plate. the so-called Mueller matrices. tal polarisation to be corrected with an accuracy of better than 1% in linear After about eight years of successful As a Nasmyth focus instrument, NACO polarisation. In addition we propose an operation of NACO and plenty of polari­ shows an instrumental polarisation observation strategy that is expected to metric observations of the Galactic Cen­ that depends strongly on the parallactic allow instrumental polarisation effects tre (GC), Sgr A*, and of many other tar­ angle of the observed source. The cause to be corrected to an accuracy of about gets, it now seemed plausible to analyse of the variable part of the IP is the fold- 0.1 %. the polarimetric mode on the basis of ing mirror M3 that is tilted at 45° (see Fig­ public ESO archive data. Long polarimet­ ure 1). While NACO itself is de-rotated ric light curves of bright sources at the with respect to the source, the orientation Polarimetric measurements can provide GC have been especially useful in investi­ of M3 and, therefore, also the orientation important information on the nature of gating the dependency of the IP on the of its IP with respect to the polarisation radiative processes, which completes the telescope position, and in determining of the observed source, changes with the picture obtained from total intensity ob- the accuracy achieved to date and attain­ parallactic angle. As shown in Figure 1, servations. In particular this is the case able for future observations. other mirrors within NAOS and CONICA for the supermassive black hole at the centre of the Milky Way and its associ­ ated variable near-infrared source Sagit­ tarius A* (Sgr A*). Here, as for other tar­ Figure 1. Optical elements of VLT UT4, NAOS and gets, the interpretation of the polarimetric CONICA and their relative orientation are sketched at the moment of the meridian transit. data depends crucially on the quality of the polarisation calibration and the knowledge of instrumental systematic errors. The majority of polarimetric obser­ vations of Sgr A* have been conducted with the near-infrared imager NAOS– CONICA (NACO) at the VLT, and we re- , #DSDBSNQ port here the results of a comprehensive analysis of the instrumental polarisation - .2 (IP) of this instrument. Full numerical 6NKK@RSNM details can be found in Witzel et al. (2010).

NACO and its polarimetric mode , , NACO is the adaptive optics near-infrared imager at VLT Unit Telescope 4 (UT4, λOK@SD 2NOSHBR Yepun), consisting of the adaptive optics

The Messenger 142 – December 2010 5 Telescopes and Instrumentation Witzel G. et al., On the Instrumental Polarisation of NAOS–CONICA

contribute as well. Since these mirrors  Figure 2. The polarisation degree of are part of the de-rotated instrument their the instrumental polarisation as pre­ dicted by the model is shown as a contribution is, however, constant. function of hour angle. Linear polarisa­  tion (in green) and total polarisation

An important effect is caused by the  (including the circular polarisation, in D red) are shown for an unpolarised HWP. A maintenance inspection revealed QD

DF source with (dashed line) and without that the actual reference system for C  (solid line) the systematic effects of the NM

the HWP position angle is turned by –6.6 @SH analyser and flat-fielding. +/– 0.2° with respect to the reference K@QHR assumed in the NACO manual. This re- /N  sults in an angle offset of –13.2° for measurements of the polarisation angle that has to be compensated for.  l l      A simulation of the polarisation degree of 'NTQ@MFKDG@ the IP as a function of hour angle (for a source at the position of Sgr A*) is shown Figure 3. Stokes Q (in blue) and U (in in Figure 2. The maximum instrumental  green) parameters as a function of hour angle for IRS16C (linear polarisa­ polarisation is about 4% (all effects, in- tion 4.6 % at PA18°, mk = 9.55). The cluding the optical elements of CONICA) solid line is the best χ2-fit of the model and drops significantly at the time of me- and the points represent observa- ridian transit. A comparison of our model tions from 2009. The fitting confirms  the time dependency of the instru-  with the observed time series of the 4  mental polarisation as predicted by Stokes parameters of bright sources at 0 the model­ and additionally allows the GC and with standard star observa­ the apparent polarisation of this star to tions confirms the model (see Figure 3). be determined. The polarisation of standard stars can be calibrated with an accuracy of better than ~ 0.5 % for the Q and U Stokes l l l     parameters. 'NTQ@MFKDG@

To our knowledge our results for the IRS16 stars at the GC are the first polari­ additional angles — with a 90°-offset with to three, mainly describing the variable metric measurements of sources within respect to the above-mentioned four part of the IP that is known with an accu­ the central parsec of the GC, since angles — intensities are measured. With racy of about 0.1% (see Figure 3). A dis­ Knacke & Capps (1977), that are inde­ this additional information Stokes param­ advantage is the alteration of the field of pendently calibrated using a method that eters that are corrected for parts of the view that results from a rotation of NACO goes beyond “boot-strapping” proce­ IP can be calculated. In particular, the as a whole as compared to the full field. dures. An example is shown in Figure 3. effects of the Wollaston and the flat-field­ The accuracy for sources of this bright­ ing can be compensated for by this strat­ ness (mk > 9.5 mag) is better than ~1% egy. Polarised emission of Sgr A* in linear polarisation and better than ~ 5° in polarisation angle for polarisation Unfortunately this calibration is not suita­ Figure 4 shows the derived total and degrees higher than ~4 %. ble to correct all instrumental effects. In polarised flux, linear polarisation and particular, the effects of reflections at polarisation position angle for Sgr A* metallic surfaces cannot be eliminated measured with NACO in June 2006. As a A strategy for high precision polarimetry completely and reference offsets as result of our analysis we can show that caused by the HWP remain uncompen­ the common boot-strapping calibration of The remaining systematic uncertainties of sated. But we can show that this method time series of Sgr A* on the basis of the about 1% in linear polarisation result from is very useful if: 1) the HWP offset is al- average foreground polarisation as deter­ insufficient knowledge of the characteris­ ready compensated for during observa­ mined by Knacke & Capps (1977) is in tics of the Wollaston prism (such as the tion; 2) the HWP is used to switch be- good agreement with our new calibration. relative transmission of the orthogonal tween the 0°/90° and 45°/135° angle The main reason is that the statistical channels) and the effects of the flat-field pairs (respectively between their corre­ errors of the photometry dominate the calibration (the calibration light for the flat sponding orthogonal pairs); and 3) NACO described instrumental effects for short field is potentially polarised intrinsically as a whole is rotated to switch by 90° for integrations of faint sources like Sgr A*. up to 1%). A common calibration method the measurement of the additional four Only at the lowest states of polarisation to circumvent problems like this is “chan­ angles. With this strategy the number of do both methods deviate significantly nel-switching”. In this method, at four parameters of our model can be reduced in polarisation angle, an effect that can

6 The Messenger 142 – December 2010 )TMD be explained by the interplay of the HWP  offset and the boot-strapping method

  that cannot correct for it. In these polari­ )X sation states a reliable estimation of the L  polarisation angle is impossible anyway, %KTW  and we can conclude that for Sgr A* in its

 bright flare phases the boot-strapping   calibration yields the same results within the statistical uncertainties as a calibra­  tion with our more elaborate polarisation  F model. Only these bright phases have

CD    been interpreted in the framework of

MFKD  the relativistic modelling (Zamaninasab et K @ l al., 2010; Eckart et al., 2006; Meyer /N l l et al., 2006) that highlights the influence   of strong gravity in the vicinity of the 4 × 106 M supermassive black hole at  A  the centre of the Milky Way.   D  QD

DF  Acknowledgements

K C 

/N  We gratefully acknowledge fruitful discussions  with Christian Hummel and the ESO User Support     Department. 3HLDLHM

References Figure 4. Light curve of Sgr A* observed with NACO in June 2006. Upper panel: total flux (black) and Eckart, A. et al. 2006, A&A, 455, 1 polarised flux (blue); middle panel: polarisation posi­ Knacke, R. F. & Capps, R. W. 1977, ApJ, 216, 271 tion angle; lower panel: linear polarisation. For the Meyer, L. et al. 2006a, A&A, 460, 15 polarimetric parameters, the results of both calibra­ Meyer, L. et al. 2006b, A&A, 458, L25 tion methods are shown: boot-strapping (red points) Witzel, G. et al. 2010, A&A, in press and the model described here (black points). Zamaninasab, M. et al. 2010, A&A, 510, A3+

Image of NACO at the Nasmyth focus of VLT UT4 (Yepun). NACO consists of the adaptive optics ­module NAOS mated with the near-infrared imager and spectrograph CONICA.

The Messenger 142 – December 2010 7 Telescopes and Instrumentation

Upgrading VIMOS

Peter Hammersley1 complete the upgrade in phases. This Following tests on the instrument in Lise Christensen 2 article describes the activities of the first November 2009 and February 2010, Hans Dekker1 upgrade and the resulting improvement in VIMOS was removed from the telescope Carlo Izzo1 performance. between May and July 2010, so that Fernando Selman1 the first stage of the upgrades could be Paul Bristow1 made. The instrument was re-commis- Pierre Bourget1 The instrument and upgrades sioned at the end of July. Making fine Roberto Castillo1 adjustments and optimising the system, Mark Downing1 VIMOS has four identical arms, each with then took a few more weeks. Nicolas Haddad1 a 7 × 8 arcminute field of view on the sky Michael Hilker1 with a gap between the fields of 2 arc- Jean-Louis Lizon1 minutes. The instrument offers three main New shutters and other measures to Christian Lucuix1 observing modes: improve reliability Vincenzo Mainieri 1 – U-, V-, B-, R-, I- and z-band imaging Steffen Mieske1 covering four fields each 7 × 8 arc- The exposure is controlled by 100 mm Claudio Reinero1 minutes in size. iris shutters manufactured by the Prontor Marina Rejkuba1 – Slit-based multi-object spectroscopy company. Due to the size and speed Chester Rojas1 with spectral resolutions from a few requirements, these shutters are subject Rubén Sánchez-Janssen1 hundred to 2500 in each of the four to considerable wear that has led to Alain Smette1 imaging fields. occasional failures. A failure occurring at Josefina Urrutia Del Rio1 – Integral-field unit (IFU) spectroscopy the end of science integration is espe- Javier Valenzuela1 with fields of view between 13 × 13 and cially frustrating because it means that Burkhard Wolff 1 54 × 54 arcseconds. the exposure is lost and must be counted as technical downtime. Analysis of night After eight years of operations, and to reports from the last two years shows 1 ESO extend its useful life, it became necessary that the technical downtime due to shut- 2 Excellence Cluster Universe, Technische to upgrade the instrument in order to ter or detector controller failures has Universität München, Germany address various issues. The first upgrade been 0.9 % on average, a non-negligible included: fraction of the total VIMOS downtime 1) replacing the shutters, which were which varied between 6 % (Q4/2009) and The high multiplex of the VLT visible worn out, and improving the reliability 10 % (Q4/2008). imager and multi-object/integral- of the instrument; field spectrometer, VIMOS, makes it a 2) replacing the CCD detectors; Since this type of shutter is no longer powerful instrument for large-scale 3) reducing the instrument flexure; commercially available, ESO has reverse- spectroscopic surveys of faint sources. 4) improving the data reduction pipeline. engineered the Prontor shutters and Following community input and recom- built ten copies. A new shutter controller mendations by ESO’s Science and was also designed, which has no dissipa- Technology Committee, it was decided tion in the open or closed condition and to upgrade the instrument in phases. has an electromagnetic braking function The first phase of the upgrade is to reduce the mechanical loads. Proto- described and included changing the types have survived lifetime tests of up to shutters, installing an active flexure 200 000 operations without degradation. compensation system, replacing the Four of these new shutters have been detectors with CCDs with a far better mounted and no technical downtime due red sensitivity and less fringing, and to shutter failures has been experienced improving the data reduction pipeline. since then.

The Mask Exchange Unit (MEU) of each VIMOS (Figure 1) is a powerful visible of the four VIMOS arms consists of (360 nm to 1000 nm) imager and four mechanisms (mask selector/gripper/ mult-object/integral-field spectrometer translator/blocker) that must be precisely mounted on VLT Unit Telescope 3 aligned for reliable operation. About (Melipal). Following a workshop on spec- half of the total technical downtime of troscopic surveys1 and recommenda- VIMOS is due to MEU failures. In the first tions by the ESO Science and Technol- quarter of 2010 we began performing ogy Committee, it was decided that daytime dry runs with the MEU whenever the instrument, which entered operation new masks were mounted. The idea in 2002, should be upgraded. In order Figure 1. VIMOS is shown on the Nasmyth platform was to detect possible failures during to miminise downtime, it was decided to of VLT UT3. daytime, when they can still be corrected.

8 The Messenger 142 – December 2010 Apart from the extra manpower to per­ 5 × 10 4 Figure 2. Raw stellar spectra taken form the daytime tests, the experience with the old (red) and new (black) detectors using the HR–red grism. The with dry runs is encouraging, although star observed with the old detector 4 × 10 4 more statistics are needed. was over a magnitude brighter that the star observed with the new detector. The “noise” in the red spectrum is 3 × 10 4 caused by the fringing which makes The new detectors the detector response vary rapidly Counts with wavelength. This fringing has al- 4 The original VIMOS detectors were 2 × 10 most disappeared with the new de­- thinned back-illuminated CCDs made by tector so that the absorption features, such as the Ca ii triplet lines, can be 4 e2v. These were cosmetically very clean 10 easily seen. and gave a good performance in the blue and visible wavelength ranges. At 0 wavelengths longer than about 700 nm, 850860 870880 however, the detectors had strong fring­ Wavelength (nm) ing, which meant that near 850 nm the quantum efficiency (QE) of the detector The active flexure compensation els when the instrument rotated. These could change by up to 40 % for changes flexures had a major impact on opera­ in wavlength of about ten nanometres, VIMOS is a large instrument, weighing tions. The spectroscopic flat and arc cali­ or a movement across the detector of a three tonnes, and suffers from 12–20 pix­ brations had to be taken immediately few pixels. This made obtaining high els of flexure between the focal plane and after the science observations to ensure quality spectroscopy very difficult in this the detector, when the instrument is ro- that the flexures affecting science and wavelength range, particularly when there tated to follow the field rotation. This flex­ calibrations were as similar as possible. was flexure present. ure causes the image to smear during Furthermore, the relative pointing of the long exposures, makes calibration more four arms between pre-image and spec­ The new detectors are also e2v devices. difficult, and reduces the accuracy of troscopic observation could change, These are the same format as the origi­- object position measurements taken in thus offsetting the sources in the slit. This nal but the silicon is more than twice the imaging mode, necessary for MOS was particularly annoying, as observers as thick. This has dramatically reduced mask production. The instrument was could never optimally position the targets the fringing so that now the maximum delivered with a passive (mechanical) in all four quadrants at the same time. change in QE is at most 2% and cannot flexure compensation system. Its perfor­ be seen at all with the LR-Red grating mance was improved by ESO to about Therefore, it was decided to install an or in imaging. Figure 2 shows two raw four pixels, but the system is not easy to active flexure compensation system stellar spectra near 850 nm taken with maintain or adjust, and in 2009 some (AFC). Two motors were placed on the the old (red) and new (black) detectors arms exhibited flexures of nearly six pix­ fold mirror of each arm, allowing the using the HR-Red grating. With the old detector the fringing is so large that none of the stellar features can be rec­ognised, and so careful reduction is required to 1 1 produce usable results ­(Scodeggio et al., 2 2009). The raw spectrum taken with the 3 new detector, however, is almost good 4 enough to use, so simplifying the reduc­ tion and improving the quality of the re- .) ag sults, especially on faint red objects. 0.5 (m

Figure 3 shows the change in imaging point zero points between the old and new ro ze detectors. The significant improvement in in the R-, I- and z-bands can be clearly e seen, however there is a slight decrease Chang 0 in sensitivity in the U-band. The data used for calculating the zero point Figure 3. The change in imaging zero point (in with the new detectors were taken just magnitudes) between before the primary mirror of Melipal the old and new detec­ was recoated, which has improved the tors is shown. The zero points by 0.05 to 0.1 magnitudes. results for each of the –0.5 U BVRIz four arms (indicated 1-4) are shown by separate Filter points.

The Messenger 142 – December 2010 9 Telescopes and Instrumentation Hammersley P. et al., Upgrading VIMOS

Figure 4. A close-up of exposures. Using the new recipes is the motors of the active mandatory for reducing data obtained compensation system. after the VIMOS CCD upgrade, and they can also be used for reducing older data. This new software, developed at ESO, is intended to replace the original set of five MOS recipes.

The new calibration recipe uses a calibra­ tion approach, based on pattern rec­ ognition, which was already applied suc­ cessfully to the FORS1/2 and EFOSC2 pipelines (Izzo et al., 2007). This greatly reduces the software maintenance work­ load as it no longer requires any prelimi­ nary optical and spectral modeling of the instrument. This approach has been adopted in order to cope with the ­me- chanical instabilities affecting any real- image of the focal plane to be displaced two pixels, the­latter degraded by hystere­ world instrument, a problem which was in X and Y on the detector. A fibre in the sis. Duringthe commissioning it was grat­ especially felt with VIMOS. Supporting focal plane, but just outside the nominal ifying to be able to see all targets centred the new VIMOS mosaic commission- imaging field, then acts as a reference in all quadrants at the same time! ing would have been impossible using source and so before an observation is the old MOS pipeline. VIMOS, with its started the image of the focal plane can four quadrants and six grisms, required be correctly positioned on the detector. The new MOS pipeline the manual recomputation of at least Any flexure during an observation is 72 spectral distortion models at each ­corrected by driving the motors using a In the new release of the pipeline data major instrument intervention. With the look-up table. With the current system reduction software (version 2.5.0, deliv­ new recipes, this is no longer necessary. the target positioning can be done with ered to the users in October 2010) two approximately one pixel accuracy, while new MOS recipes have been added; one The new recipes also significantly im­­ the registration between science and for processing the calibration frames, prove the accuracy of the wavelength cal­ ­calibration observations is approximately and another for reducing the scientific ibration, the quality of the sky sub­traction,

Figure 5. The image 27ǥ shows the galaxy NGC 6754 observed with the HST/NICMOS in the J-band. From the two IFU pointings the oxygen abundance was determined in the regions corresponding 9.4 to where four SNe exploded (shown as open circles).

9.1 H) O/ + log(

8.8 12

8.5

10 The Messenger 142 – December 2010 7 Figure 6. The VIMOS colour maps on the right-hand side of spectrum of one single Figure 5 illustrate this abundance, where Hα spaxel of the IFU obser­ 6 vation of NGC 6754. the dark blue colour corresponds to solar metallicities, green and red to super-solar 5 metallicities, and violet and black to sub-

x solar metallicities. Figure 6 shows an

flu 4 [NII] λ6583 extracted spectrum from a single spaxel

ve identifying the lines used, but also dem­ 3 onstrating fewer systematic sky subtrac­ Relati tion residuals beyond 700 nm, thanks to H 2 β the absence of detector fringes. The sig­ [O III] λ5007 nal-to-noise ratio for continuum emis- 1 sion is 50 % higher, while in the region of strong sky emission lines the improve­ 0 ment is a factor of two, relative to the old 5000 6000 7000 8000 9000 detectors. Wavelength (Å) the optimal extraction of the detected after any intervention that causes a dis­ Next steps objects, and the sky fringing correction. placement of the image of the focal plane The new pipeline and the manual can be on the detector, takes a minimum of a The first stage of the upgrade has sig­ downloaded 2. week. During the commissioning, which nificantly improved the performance was also a time to become familiar with of VIMOS in the red and reduced the the newly installed hardware, we had amount of flexure to values close to the The commissioning to re­­peat this re-characterisation a few initial specifications of the instrument. times. Although this led to a loss of Even so, further improvements are being The commissioning of an instrument as ­efficiency during the first two months of planned for 2011 and 2012: complex as VIMOS needs to be carefully ­service observing, the stability achieved 1. Replace the high resolution blue grism planned. There are many modes, and the in the coefficients of the mask-to-CCD by a volume phase holographic grating tree of dependencies has become even matrix are now outstanding, with daily with almost double the efficiency. more complex with the introduction of the positioning variations well below a pixel. 2. Improve the reliability and accuracy AFC system. The AFC reference source of the grism placement, mask insertion is at the edge of the field and so focusing and focus mechanisms. is now even more critical if defocus is Performance validation 3. Increase the speed of the grism not to cause the centroids to move. The exchange unit (expected increase in conversion from pixel coordinates to milli­ As part of the commissioning, perfor­ overall operational efficiency is 2%). metres at the mask plane, is now tied mance validation observations were 4. Eliminate the need to perform pre- to the coordinates of the AFC reference made to help characterise the MOS and imaging with VIMOS before masks can pixel. Any change to the reference pixel IFU modes and provide a taste of the be made. Potentially this could in­ means that the mask-to-CCD matrix performance after the upgrade. Figures 5 crease operational efficiency by 10 to must be re-determined. and 6 show the results from the IFU 15 %. observations. Figure 5 shows the HST/ VIMOS has three reference systems NICMOS J-band image of NGC 6754 After these changes we expect to have an which must be aligned to a few pixels in where four supernovae were observed instrument with a much improved ­stability, 4000; the detector xy in pixels, the focal between 1998 and 2005. Two belonged reliability, and efficiency that will permit plane XY in mm and the spectral dis­ to massive stars that exploded in core more ambitious science to be attempted. persion direction. This last alignment is collapse events (Type ii SNe) and the particularly important for multiplexed ob­ other two to white dwarf star explosions servations with the low resolution grisms, (Type i a SNe). NGC 6754 is dominated by References as the reduction pipelines assume that numerous Hii regions and widespread Izzo, C., Jung, Y. & Ballester, P. 2007, in The 2007 the contaminating zero or second orders star formation over the face of the galaxy. ESO Instrument Calibration Workshop, 191 are well aligned with the first order spec­ The strong optical emission lines in the Pettini, M. & Pagel, B. E. J. 2004, MNRAS, 348, 59 trum (we thank Luigi Guzzo, Bianca spectra at the two VIMOS–IFU pointings Scodeggio, M. et al. 2009, Messenger, 135, 13 Garilli, and Marco Scodeggio for strongly allows the oxygen abundance in the re- insisting on this point). There are four gions where the SNe exploded to be Links arms, and six grisms per arm, each of derived, as well as in other regions of the which needs to be aligned, and so realign­ galaxy. The oxygen abundance was cal­ 1 Conference web page including presentations: www.eso.org/sci/meetings/ssw2009/ ment of all grisms is a lengthy process. culated from the H /[N ii] ratio using the α 2 VIMOS reduction pipeline available at: The re-characterisation of the instrument calibration of Pettini & Pagel (2004). The http://www.eso.org/pipelines

The Messenger 142 – December 2010 11 Telescopes and Instrumentation

Progress on the VLT Adaptive Optics Facility

Robin Arsenault1 MUSE. These modules implement inno- improve the ensquared energy in the Pierre-Yves Madec1 vative correction modes for seeing K-band by a factor of two over the Jérôme Paufique1 improvement through ground layer 7.5-arcminute field of view of HAWK-I. Stefan Ströbele1 adaptive optics and, for high Strehl ratio The four LGS WFSs are mounted on Jean-Francois Pirard1 performance, laser tomography adap- a rotating bearing inside the Hawk-I hous­ Élise Vernet1 tive ­correction. The performance of ing, while a natural guide star tilt sensor Wolfgang Hackenberg1 these modes will be tested in Europe picks up a star in a ring outside the Norbert Hubin1 with a custom test bench called ASSIST. Hawk-I field of view. An additional natural Liselotte Jochum1 The project has completed its final guide star WFS is used for calibration Harald Kuntschner1 design phase and concluded an intense and maintenance operations of the Andreas Glindemann1 phase of procurement; the year 2011 deformable secondary mirror (DSM). Paola Amico1 will see the beginning of assembly, inte- Miska Lelouarn1 gration and tests. The GALACSI and GRAAL real-time Johann Kolb1 computers share identical SPARTA hard­ Sébastien Tordo1 ware architecture and most software Robert Donaldson1 Overview ­features. SPARTA is a standard platform Christian Sönke1 for real-time applications developed by Domenico Bonaccini Calia1 The ESO VLT Adaptive Optics Facility ESO’s Adaptive Optics Department and Ralf Conzelmann1 (AOF) is upgrading one of the Unit Tele­ uses a hybrid architecture of a field pro­ Bernard Delabre1 scopes (UTs) with a new secondary grammable gate array, digital signal pro­ Mario Kiekebusch1 ­mirror (M2) unit. This unit includes an cessor and central processing unit. It is Philippe Duhoux1 adaptive secondary mirror (with 1170 ac­­ designed to be flexible by offering sophis­ Ivan Guidolin1 tuators), four laser guide stars (LGS) ticated software control and post-pro­ Marco Quattri1 formed by four 22 W sodium beacons cessing features while accommodating Ronald Guzman1 launched from the telescope centrepiece various combinations (numbers) of WFSs Bernard Buzzoni1 and a dedicated adaptive optics (AO) and deformable mirrors. All WFS cam- Mauro Comin1 instrument park to provide users with eras and tilt sensor cameras are identical Christophe Dupuy1 optimised AO correction modes. The two and use as detectors e2v CCD220 Jutta Quentin1 AO components are GALACSI, the AO devices controlled by the New General Jean-Louis Lizon1 module for MUSE that will provide ground Controller (NGC; Baade et al., 2009) Armin Silber1 layer adaptive optics (GLAO) correction developed by the ESO Optical Detector Paul Jolly1 and a laser tomography correction for Team. Antonio Manescau1 MUSE’s narrow-field mode for high Strehl Peter Hammersley1 ratio correction in the visible spectral Javier Reyes1 range, and GRAAL, the AO module for Status of the project Andreas Jost1 HAWK-I, which will provide only ground Michel Duchateau1 layer adaptive optics correction. The GALACSI Volker Heinz1 Adaptive Secondary Simulator and The GALACSI AO module design and Clémentine Bechet1 Instrument Testbed (ASSIST) will be used manufacture is led by the ESO AO Remko Stuik 2 for a complete test phase of the AOF in Department with the support of other Europe. This project constitutes a major ESO divisions (the Technology Division, stepping stone towards the European Integration Department of the Instru­ 1 ESO Extremely Large Telescope (E-ELT). mentation Division and Software Devel­ 2 Huygens Laboratory, University of opment Division). The main mechanical Leiden, the Netherlands GALACSI is a module mounted at the structure and optics have been out­ Nasmyth focus; it contains four LGS sourced to industry and off-the-shelf wavefront sensors (WFS), a natural star components are being purchased for the The Very Large Telescope (VLT) Adap- tilt sensor and an infrared low-order sen­ electro-mechanisms. The control soft­ tive Optics Facility is a project that sor. The laser guide stars and WFSs can ware is being developed in-house and will transform one of the VLT’s Unit Tel- be tuned to two different fields of view: a integration will be undertaken by the Inte­ escopes into an adaptive telescope 1-arcminute field of view for GLAO cor­ gration Department. The mechanisms that includes a deformable mirror in its rection, producing a gain of two in ens­ are controlled via the remote motor con­ optical train. For this purpose the sec- quared energy at 750 nm; and a 7.5-arc­ troller switching system based on the ondary mirror is to be replaced by a thin second field of view for laser tomography controller area network protocol. The final shell deformable mirror; it will be pos­ correction, providing a high Strehl ratio in design review was held in June 2009 sible to launch four laser guide stars the visible (5% at 650 nm) on-axis. and several important procurements were from the centrepiece and two adaptive launched at that stage. GALACSI’s main optics modules are being developed GRAAL is similar in design to GALACSI, structure, manufactured by Bossenkool to feed the instruments HAWK-I and but only applies a GLAO correction to (the Netherlands), has gone through final

12 The Messenger 142 – December 2010 Figure 2. Several NFM Dichr. unit Beam Splitter unit smaller mechanical components of To MUSE GALACSI’s optical Comm. Camera train are shown. Field Selector LGS WFS assembly

To WFS

Figure 1. GALACSI’s main mechanical structure is shown during acceptance at Bossenkool (the Netherlands) in September 2010. Calibration unit To vis TT sensor acceptance and was delivered to ESO in October 2010; it is shown in Figure 1. from VLT CU mirror The optics is being manufactured by SESO (France) and delivery is expected in December 2010. Several smaller mechanical components have already is also being manufactured by SESO and stand-alone mode should start by late been received and are being tested and delivery is expected in December 2010. 2011 to prepare for a system test phase some are shown in Figure 2. The field Several software modules are being with the DSM on ASSIST in 2012. selector is a critical key system and the developed in common for GRAAL and positioning stages manufactured by GALACSI and there are synergies be­­ Deformable secondary mirror Physik Instrumente have been received tween both. During summer 2010, a VLT The deformable secondary mirror is being and are currently being tested. The elec­ standard software module was deliv- outsourced to MicroGate and ADS tronic control is progressing steadily ered to NTE SENER S.A. with the corre­ (Italy) and is the heart of the Adaptive and several VME racks are complete and sponding hardware to control the co- Optics Facility. This contract completed have been integrated into the cabinets. rotator (consisting of an internal bearing its final design review in December and drive system to rotate the four LGS 2007 and assembly started at the end of The assembly and integration of GALACSI WFSs) and is being used by NTE SENER 2009. Several key components have started in the ESO integration hall dur- S.A. for their integration and testing pro­ been received by the suppliers and work ing the autumn of 2010 and will continue cess, as shown in Figure 3. is progressing along two parallel paths throughout 2011. The stand-alone test at MicroGate, where the electronics and for the GALACSI module (without the When the GRAAL main structure is deliv­ software are being developed, and at DSM) will begin in 2011 to validate the ered to Garching early in 2011, the optics whole assembly without the other major will be integrated with the rest of the Figure 3. GRAAL’s main structure is shown being subsystems of the AOF. The complete ­electronics and software. GRAAL tests in integrated and tested at NTE SENER S.A. in Spain. system test phase for GALACSI with the DSM on ASSIST is expected to start in early in 2013 after GRAAL has completed its corresponding phase.

GRAAL The GRAAL module is also led by the AO Department with support from the same departments as GALACSI, but with one main difference: a major contract has been issued for the important task of manufacturing and integrating the main structure and opto-mechanics, including all stages (provided by ESO) and the WFS co-rotator. This contract is being undertaken by NTE SENER S.A. (Spain) and will be completed in early 2011. This scheme will give the integration activities a strong start in Garching. The optics

The Messenger 142 – December 2010 13 Telescopes and Instrumentation Arsenault R. et al., Progress on the VLT Adaptive Optics Facility

Figure 4. Part of the consist of 1178 nm infrared fibre Raman DSM reference body lasers, doubled in frequency by second made by SESO in France (see the image harmonic generation to 589 nm (see on p. 4, upper, for the Bonaccini Calia et al. [2010] for details); whole DSM). The intri­ Figure 7 shows the test laser setups. cate rib structure is These lasers have very few components, ­literally sculpted from a monolithic block of require minimal optical alignment and Zerodur. The 1170 circu­ occupy a small volume, all contributing lar holes allow passage to a greater reliability. This represents for the actuators. This a major milestone in terms of simplicity, process, known as light­ weighting, reduces the compactness and improved reliability. mass by 80 % to a mere The current LGS system uses the 35 kg. ­PARSEC laser, which delivers 5 W of Na power and sits on a large optical bench located in the laser cleanroom below the Nasmyth platform of UT4. In contrast the TOPTICA 22W laser fits in a volume ADS where the electro-mechanics is DSM. The next step is a milestone in mid- of half a cubic metre (excluding the elec­ being assembled and manufactured. The 2011 to review the progress of the inte­ tronics cabinets)! main highlight of 2010 has been the gration. Delivery of the DSM to Garching ­delivery of the reference body, a Zerodur is planned for early 2012, when the TNO Science & Industry (the Nether- 1.1-metre lightweighted optical compo­ ­system test phase in Europe will begin. lands) holds the second important con­ nent that provides the reference surface tract. They will finalise the reference for the thin shell mirror, shown in Figure 4 The Four Laser Guide Star Facility design provided by ESO and manufac­ (and featured in an ESO announcement1). The Four Laser Guide Star Facility (4LGSF) ture the optical tube assembly, which project includes two important con- is the telescope part of the launch tele­ SAGEM (France) is continuing its efforts tracts contributing to the design and scope system (see Figure 8). This is on the thin shell contracted to them manufacturing of this system. TOPTICA based on a two-lens design, the largest by MicroGate. The convex side has been Photonics, a company located near being 40 cm in diameter while a 30 cm polished successfully to specifications Munich, is responsible for the laser de­ beam is launched into the atmosphere. and the thinning process from the back­ sign and manufacture. TOPTICA has A 45-degree mirror steers the beam and side is starting. Note that SAGEM is been selected as the preferred supplier allows the laser guide star to be pointed ­proceeding carefully and made a detailed after an initial call for tender and a com­ to the appropriate off-axis distance. analysis of the thinning process before petitive preliminary design with another finalising the details. Figure 5 shows the supplier. The manufacturing contract The ESO 4LGSF team is focusing on integration and service stand for the was signed in June 2010. A final design the rest of the laser system design: beam DSM and M2 hub. Figure 6 shows some review will be conducted in June 2011 diagnostics and jitter control for fast of the electronic control boards for the and six months later a pre-production beam pointing; optical and mechanical unit will be ready so that all the laser design; electronic and software design; Figure 5. Integration and service stand for the DSM. specifications can be verified. Then and analysis of the whole system. Inter­ The M2 hub, including the DSM, is mounted inside. the four units will be produced. They facing the 4LGSF with the UT and the

Figure 6. Electronic control boards for digital signal processing for the DSM. Each board controls 16 channels.

14 The Messenger 142 – December 2010 Figure 7. Left: The ESO Raman fibre laser used to test laser optics coatings for the launch telescope. Right: The single arm Raman fibre amplifier from MPB/TOPTICA showing the required 37.4 W of infra­ red 1178 nm power for the proper Na wavelength output power (image from April 2010).

unit is expected for the summer of 2011 and the final design of the laser will be reviewed in mid-2011.

ASSIST The ASSIST test bench is a key com­ ponent of the Adaptive Optics Facility. It will allow full characterisation of the Figure 8. The optical AOF in Europe before installation and Lens L2 & mount tube assembly being com­missioning in Paranal. It also offers developed by TNO Science & Industry. ­stable, known and reproducible condi­ Main tube Together with the beam tions of turbulence (simulated by phase control and diagnostics screens) to establish whether the whole system this assembly assembly (wavefront sensors, real-time makes up the launch Optical computer, deformable mirror and the base block telescope system for the 4LGSF. Four similar algorithms implemented) meet the perfor­ units are mounted on mance specifications. Field selector the centrepiece of the mechanism UT to launch the four laser guide stars at The most striking feature of ASSIST is 90 km altitude. that it requires a large 1.7-metre main Lens L1 & quarter- wave plate mirror to feed the light onto the M2 con­ vex deformable mirror. This mirror is manufactured by AMOS (Belgium) and is in the last stages of aspheric polishing (see Figure 9). Acceptance is planned for late 2010.

The bench consists of a tower with the 1.7-metre mirror at the base and the new generation M2 unit (hosting the DSM) on top (see Figure 10). A system of 45-degree mirrors will feed artificial sources from the side. A star simulator and turbulence generator feeds ASSIST with sources simulating natural and laser guide stars (with different focus positions) and three phase screens located at different altitudes. This setup allows realistic testing of the AO modules Figure 9. The ASSIST test bench (left) with the new lasers as components and is taken very and complete testing of all the functio­ generation M2 unit on top of the structure. The large seriously by the team. A consultant has nalities. ASSIST will be delivered to 1.7-metre mirror is at the bottom (hidden by the light covers in this illustration) and the Nasmyth adaptor/ been contracted to provide support Garching in spring 2011 for integration in rotator test bench can be seen on the right-hand to the safety analysis and much has been the large tower with the mirrors well in side to allow the AO modules to be mounted. On the learnt during the design of the current ad­­vance of the arrival of the DSM, planned right is the secondary mirror (AM2), a 140 mm 4LGSF system. for early 2012. ­convex aspherical mirror manufactured by ASTRON, that completes the main tower optics of ASSIST. The final design documentation for the Unit Telescope upgrade 4LGSF is complete and the final design In April 2008 it was recognised by the mechanical design of the lasers and opti­ review will take place in January 2011. project and upper management that the cal tube assembly interfaces also re­ Procurement will then start for the opto- scope of modifications required to the UT quires the team’s full attention. Safety is mechanical and electronic system that will receive the AOF were substantial important when designing with 22 W assembly. A first optical tube assembly and needed detailed planning and careful

The Messenger 142 – December 2010 15 Telescopes and Instrumentation Arsenault R. et al., Progress on the VLT Adaptive Optics Facility

Figure 10. The main MUSE narrow-field mode of GALACSI. 1.7-metre mirror of The ESO Optical Detector Team is us- ASSIST is shown at AMOS in Belgium ing the fast readout OCam (developed by during the last step of LAOG, LAM and OHP; see Release aspherical polishing. eso09222) to develop an NGC controller for the e2v CCD220 detector that will be used for the WFS cameras. Fifteen units have been ordered for SPHERE and the AOF.

Project outlook

The year 2011 will be an exciting one dur­ ing which all the AOF project subsystems will start assembly and testing. The first science thin shell mirror and the GRAAL main assembly should be delivered dur­ ing the first quarter of the year. GALACSI integration will already be underway and preliminary tests of the two AO module subsystems will take place in the course of the year. ASSIST will be delivered to Garching and integrated during the sec­ ond half of 2011.

The AOF system test phase will start when the DSM is delivered in the first quarter of 2012 and will proceed with optical testing of the DSM, full system tests of GRAAL and then GALACSI. This Figure 11. Left: Concept of a Nasmyth platform distance of GALACSI–MUSE; defi­nition of will last for slightly more than one year. extension for the 4LGSF electronic cabinets and all cable (for power, signals, fibres, etc.) The 4LGSF integration and test proceeds heat exchangers below the existing platform. Right: One of the two concepts for a deployable M2 main­ routing in the UT; development and imple­ in parallel to these activities and the first tenance platform to enable quick maintenance work mentation of a new, easier to deploy, shipments to Paranal and the beginning on the new generation M2 unit. M2 maintenance platform (see Figure 11 of the commissioning activities will take right); and specification of all the required place during the second half of 2013. implementation. A specific branch of the assembly, integration and testing facilities This shipment will trigger intense com­ AOF project was created to manage this needed in Paranal. missioning activities involving the whole effort. At the AOF system final design AOF team until the end of 2014, which review, held in April 2010, a large number Complementary systems is the project goal for provisional accept­ of documents focused on this aspect and There are a few important components ance Chile. received the approval of the review board of the AOF that are being developed and the Paranal management. This re- as sub-contracts to other firms or to ESO view constituted the kick-off for this sub- departments. The real-time computer References project. The basis for the ­definition of the SPARTA for GRAAL and GALACSI is Arsenault, R. et al. 2010, Proc. SPIE, 7736, 0L requirements is all the non-conformances managed by a sub-group of the AO De­ Baade, D. et al. 2009, The Messenger, 136, 20 identified in the Interface Control Docu­ partment. The two identical real-time Bonaccini Calia, D. et al. 2010, The Messenger, ment between the AOF and the VLT. This computers will be delivered to the AOF 139, 12 sub-project will implement: mechanical after SPHERE is delivered. SPHERE Feautrier, P. et al. 2010, Proc. SPIE, 7736, 0Z Paufique, J. et al. 2010, Proc. SPIE, 7736, 1P interfaces between the UT and the is a single conjugate AO system using a Stuik, R. et al. 2010, Proc. SPIE, 7736, 3M 4LGSF electronics cabinets and launch natural guide star on axis; the required telescopes on the UT centrepiece (see functionalities will be re-used by the Figure 11 left); overhaul of the cooling maintenance mode of GRAAL (for the Links system and electrical supply; implemen­ natural guide star on-axis). The GLAO 1 Deformable secondary mirror: http://www.eso.org/ tation of platform extensions to accom­ functionalities will build on this for the public/announcements/ann1056/ modate all the electronics cabinets; GLAO mode of GRAAL and GALACSI. 2 OCam: http://www.eso.org/public/images/ ­modification to the Nasmyth guider arm These aspects will be completed by laser eso0922b/ to accommodate the longer back focal tomography AO functionalities for the

16 The Messenger 142 – December 2010 Telescopes and Instrumentation

ALMA Status and Progress towards Early Science

Leonardo Testi1 team is working to obtain science dem­ mentation section page, p. 4); the remain­ Richard Hills 2 onstration data with the eight-antenna der are partly equipped and under test or Robert Laing1 interferometer at the high site and is in the process of being outfitted with Stefano Stanghellini1 focusing on refining the calibration plan, ALMA front-end and back-end systems Wolfgang Wild1 in addition to continuing the commis­ at the Operations Support Facility (OSF). sioning activities on the equipment that Many more antennas are in various is constantly being delivered. stages of assembly in the contractor 1 ESO camps at OSF. Among these are the first 2 Joint ALMA Observatory On the hardware construction side, there European antennas from the AEM con­ has been steady progress throughout sortium; five of these have been fully the year thanks to the tremendous effort assembled at the time of writing (see Fig­ We report on the status and progress from the project teams in all regions ure 3) and the first ones are undergoing of the ALMA project and the expected and the work of the Assembly Integration performance tests and tuning with the timeline and capabilities for Early Sci- and Verification group in Chile. In par­ goal of being delivered to the ALMA pro­ ence. Over the past year, the progress ticular, focusing on deliverables from the ject in early 2011. on ALMA construction and on the com- European side, the ALMA front-end missioning activities has been huge. cartridges for Band 7 (275–373 GHz, pro­ Initial results of the ­all-sky pointing tests At the time of writing the observatory is duced at IRAM in France) and Band 9 have been excellent with the root mean progressing on the initial phases of sci- (602–720 GHz, produced by NOVA in square on pointing accuracy below 1 arc­ ence verification and preparing to open the Netherlands) as well as the water second (see Figure 4 for an example of a to external users to begin Early Science vapour radiometers (produced by the pointing run). The surface accuracy of the observations. Swedish company Omnisys) are in a very European antenna tested so far is also advanced stage of production and more excellent. The panels are mounted and than half of the total number of units aligned in the assembly procedures to a The year 2010 has seen extraordinary have been delivered to the project. The good initial accuracy. Preliminary tests progress in the ALMA project. At the test data shown in Figure 2 were ob­ on the first antenna have shown that, fol­ ­be­­­ginning of the year, with three fully tained with the Band 7 and Band 9 car­ lowing a partial initial cycle of holography equipped antennas working as an inter­ tridges. Phase correction using the water measurements and panel adjustments, ferometer, Commissioning and Sci- vapour radiometers (see Nikolic et al., the surface can be brought within the ence Verification (CSV) activities officially 2009) has been demonstrated as part of ALMA specifications (see example in Fig­ started at the Array Operations Site commissioning; this represents a key ure 5). Testing of many aspects, such as (AOS) on the Chajnantor plateau at an step towards validating the ALMA calibra­ the dependence of the antenna perfor­ elevation of 5000 metres (Testi, 2010). tion strategy for high frequency and high mance on environmental conditions and Under the leadership of Richard Hills and angular resolution observations. of the final figure and stability of the sur­ Alison Peck at the Joint ALMA Office, face ac­­curacy, are starting at the time of the CSV team has been steadily working Naturally, the most visible pieces of writing. to test and improve the performance ­hardware of the ALMA system are the of the whole ALMA system. This process antennas. ALMA has provisionally Figure 1. ALMA test data obtained in Bands 3 and 9 has culminated in the production of the accepted fourteen antennas. Eight anten­ on the nearby star-forming galaxy NGC 253. The large image is a VISTA infrared image of NGC 253 first ALMA test images in the second part nas are fully equipped and working as (see ESO PR 1025). The data shown is the CO(1–0) of the year (see Figure 1 for an early an interferometer at the high site (see the integrated emission; at right the ­CO(6–5) integrated example). At the time of writing the CSV image on the Telescopes and Instru­ emission contours are overlaid on the 0.45 mm con­ tinuum emission.

+, ".)l(MSDFQ@SDC(MSDMRHSX -&"!@MCBNMSHMTTL ".)lBNMSNTQR ǥ

ǥ ǥ

Credit: ALMA CSV team S@ lŒࣳǥ @ S@ ǥ @ C C ǥ RS S D ǥ R ǥ S D S ǥ ǥ , +, + ǥ ǥ  ǥ lŒࣳǥ

GLR R R R R GLR  R  R  R  R  R 

The Messenger 142 – December 2010 17 Telescopes and Instrumentation Testi L. et al., ALMA Status and Progress towards Early Science

IRAS4B contin & hv CO

20ǥ

16ǥ

Credit: ALMA CSV team a 12ǥ t

a consortium ESO/AEM Credit: d 08ǥ st e t 04ǥ A 31º13ࣳ00ǥ M AL 56ǥ

52ǥ 03h29m13s.5 13s.0 12s.5 12s.0 11s.5 Right Ascension (J2000)

Figure 2. ALMA test data in Band 7 of the molecular outflow in the star-forming region NGC 1333. The red and blue contours show the high velocity ­CO(3–2) emission while the 0.85 mm continuum emission is shown as greyscale.

Figure 3. The AEM antenna camp at the OSF. Five antennas are fully assembled while a sixth is inside the temporary shelter on the right ready for the installation of the main reflector assembly; parts of the seventh antenna are also being assembled.

C2 C$ C5 Figure 4. Example of all-sky optical    pointing session for the first AEM antenna. The top panels show the pointing errors as a function of ­azimuth and elevation of the observed stars. The bottom­ right circular --$6--$6--$6 ­diagram shows the position of the observed stars on the sky. The bottom left cir­cular panel shows the point- l l l ing error distribution, which in this case corresponds to an all sky rms accuracy of 0.8 arcseconds. C2 C9 ,   

9 9   --$6

l l l

4 - 7 # ǥ

1 2 9   + 6   @QBRDB

1   @QBRDB AHMR

+@RSNARDQU@SHNM.BSNADQ

18 The Messenger 142 – December 2010 Credit: ESO/AEM consortium ESO/AEM Credit:

Figure 6. Example results of a holography run after tuning the position of approximately half of the pan­ els on the AEM dish. In this particular case the sur­ face was found to be set to approximately 15 μm rms.

Figure 5. Adjusting panels on the AEM antennas standards expected for ALMA data and spectral resolution correlator modes. using the manual panel adjuster tools. Note that this run science operations at the same time During the first year of operations, scien­ operation can be performed walking on the reflector surface when the antenna is stowed; the hole on the as construction and commissioning. For tific observations will be limited to a frac­ dish surface to the rear of the image is used for the these reasons the ALMA science advi­ tion of the total available time, not ex­ optical pointing telescope. sory committees have all recommended ceeding 30 %, as the first priority of the that the main focus remains on the full project will be to work on completing and ALMA construction and operations and commissioning the full ALMA system. that the Early Science capabilities are focused on a limited set of well-tested A series of workshops, schools and During the autumn of 2010, the ALMA modes. ­tutorials have been and are being organ­ project has been going through a series ised to prepare European astronomers of external reviews that have focused The announcement of opportunity for for ALMA. The European ALMA Regional on the progress and current results of the Early Science observations will be re­ Centre (ARC) and its nodes located in CSV activities and the readiness for oper­ leased in the first quarter of 2011 with the several European countries have put in ations, culminating in the ALMA Annual deadline for proposal submission before place a strong effort to support the com­ External Review. These reviews scruti­ the summer. The precise capabilities that munity uptake of Early Science. On 6 nised all the aspects of the ALMA project will be available for Early Science will be and 7 April 2011, the European ARC will and provided advice to the ALMA Direc­ announced as part of the call for propos­ organise a tutorial at ESO Headquarters tor and the ALMA Board on the status als, but it is likely that the array will be in Garching focused on the use of the and progress of ALMA construction. The offered in two compact configurations ALMA software for Early Science (see the main focus of the reviews has been to of 16 antennas (with the maximum base­ full announcement on page 49). In par­ assess the readiness of ALMA for Early line likely to be of the order of 250 metres). ticular, the tutorials will focus on the use Science. The conclusion of the reviews is The limited instantaneous coverage of of the ALMA observing tool, which will that hardware delivery, system verifica- the uv-plane will imply that most projects be used by ALMA users for the prepara­ tion and commissioning, as well as the will require Earth rotation synthesis to tion of observing proposals and to set up deployment of all the procedures for the achieve good imaging capabilities during the approved projects for observation at start of science operations, are all pro­ Early Science. The available observing the telescope. gressing at the pace required to start modes will be limited to single field inter­ Early Science observations in the second ferometry using Bands 3, 6, 7 or 9 half of 2011. Nevertheless, the opening (84–116 GHz, 211–275 GHz, 275–373 GHz, References of the observatory to external users will 602–720 GHz, respectively), full polari­ Nikolic, B. et al. 2009, The Messenger, 131, 14 result in additional challenges to meet the sation capabilities and a range of single Testi, L. 2010, The Messenger, 139, 52

The Messenger 142 – December 2010 19 the region. from illuminated molecular dusty clouds encircle nebulosity reflection and clouds Dark Coronet. the stars, young of cluster embedded an by rounded sur is which star), Ae/Be Herbig (an itself Australis R Coronae is centre the in star The combined. Images through B telescope. 2.2-metre MPG/ESO the with taken Australis RCoronae region formation star mass intermediate—low nearby the of image Colour Astronomical Science -, -, V - and R - and -band filters were were filters -band - Astronomical Science

Precise Modelling of Telluric Features in Astronomical Spectra

Andreas Seifahrt1,2,3 transmission losses in the near- and mid- and, in these cases, is automatically pro­ Hans Ulrich Käufl 3 infrared, often hampering the observation vided for service mode observations. Günther Zängl4,5 of important astrophysical lines. Since telluric lines generally do not scale Jacob Bean2,6 linearly with airmass and observing Matthew Richter1 While the Earth’s atmosphere absorbs ­conditions are time variable, it is neces­ Ralf Siebenmorgen 3 light from astronomical sources at a large sary to observe a telluric standard star at number of frequencies from the ultra- the same airmass and close in time to violet to radio wavelengths, it also emits the science target in the same instrumen­ 1 University of California at Davis, USA light in the same transitions, radiating tal setup. Special software tools at the 2 Institut für Astrophysik, Göttingen, its thermal energy into space. At wave­ telescope allow the efficient selection of Germany lengths longer than about 2300 nm, the an appropriate standard star. 3 ESO emission originating from the Earth’s 4 Meteorologisches Institut der LMU, atmosphere competes with the signal Nevertheless, using standard stars as München, Germany from astronomical sources. The break- empirical calibrators has several disad­ 5 Deutscher Wetterdienst, Offenbach, even point depends on many factors, vantages. The observation of standard Germany such as the spatial resolution, slit width, stars is time-consuming, especially when 6 Harvard–Smithsonian Center for and detector characteristics of the science targets are bright or high signal- Astrophysics, Cambridge, USA ­spectrograph, but also on the height of to-noise requirements are to be met. the observing site. Since the atmos- For the brightest standard stars, the time pheric emission is spatially extended, it required for telescope and instrument Ground-based astronomical observa- affects the spectrum of the astronomical presets ultimately limits the efficiency of tions suffer from the disturbing effects target similarly to the nearby sky and this method. Moreover, on instruments of the Earth’s atmosphere. Oxygen, can thus be removed by beam-switching like CRIRES, the instrumental profile water vapour and a number of atmos- (nodding) techniques or by fitting and depends on the performance of an adap­ pheric trace gases absorb and emit subtracting the two-dimensional signal tive optics system, and thus, on the light at discrete frequencies, shaping along the slit. observing conditions and source bright­ observing bands in the near- and ness. Changes in the instrumental pro- mid-infrared and leaving their finger- file lead to changes in the line shape of prints — telluric absorption and emis- Empirical calibration unresolved spectral lines. In these cases sion lines — in astronomical spectra. standard stars rarely provide a perfect The standard approach of removing the Removing the telluric absorption features match to the science target, effectively absorption lines is to observe a telluric is more difficult since their exact signa­ limiting the precision with which telluric standard star: a time-consuming and ture is only imprinted in the source spec­ features can be removed. often imperfect solution. Alternatively, trum. Traditionally, telluric absorption the spectral features of the Earth’s ­features are removed by observing a so- Other shortcomings originate from the atmosphere can be modelled using a called telluric standard star before or intrinsic stellar features of the standard radiative transfer code, often deliver- after the observation of the science target stars. Hydrogen and helium lines in ing a satisfying solution that removes at the same airmass, subsequently divid­ early-type standards cannot be perfectly these features without additional obser- ing the spectrum of the science target removed and thus affect the line profiles vations. In addition the model also by that of the telluric standard. Typically, of the same species in the spectrum provides a precise wavelength solution early-type stars with spectral types rang­ of the science target. A further complica­ and an instrumental profile. ing from early B to late A are used for this tion in removing these lines occurs purpose, since they exhibit rather fea­ when the wavelength coverage of the tureless spectra, except for strong hydro­ spectrum is shorter than the line widths The Earth’s atmosphere consists of a gen lines. Since these stars are often of broad hydrogen lines, which can eas- rich gas mixture. While its main constitu­ fast rotators, other weak stellar features, ily be the case for CRIRES. Early-type

ent, nitrogen (N2), does not exhibit any for example helium lines, are further stars can also exhibit other spectral fea­ rotational–vibrational transitions, most ­suppressed. Alternatively, early-to-mid tures, such as oxygen or carbon lines other molecules do. Strong absorption G-type stars can serve as telluric stand­ in the near-infrared, often as emission

line systems from water vapour (H2O), ards as well, since high resolution Fourier features originating under non-local ther­ carbon dioxide (CO2) and ozone (O3) transform spectra (FTS) of the Sun pro­ modynamic equilibrium (non-LTE) con­ shape the well-known photometric band­ vide the necessary template to compen­ ditions. Similarly, mismatches in the line passes in the near- and mid-infrared. sate for their intrinsic stellar features. depths of solar-type standard stars

In addition, oxygen (O2) shows strong ab­ The observation of telluric standard stars with the solar FTS atlas, attributable to sorption bands in the red optical; other is a standard procedure and commonly abundance differences or deviating effec­

molecules such as nitrous oxide (N2O), employed for all spectrographs operating tive temperatures, can leave residuals carbon monoxide (CO), or methane (CH4) on ESO telescopes. It is often part of of the standard star’s intrinsic features in contribute noticeably to the atmospheric the calibration plan for the instrument the final spectrum.

The Messenger 142 – December 2010 21 Astronomical Science Seifahrt A. et al., Precise Modelling of Telluric Features in Astronomical Spectra

Modelling telluric features     Alternatively, the telluric absorption spec­ trum can be synthesised, using a radia­   TW tive transfer code in combination with a Ck layered model of the Earth’s atmosphere   and a database containing the transition   data for all molecules in consideration. NQL@KHRD We have successfully used the program   LBLRTM (Clough et al., 1992) for this   ­purpose. LBLRTM is a non-commercial layer-by-layer radiative transfer code          ­tailored to produce telluric spectra under various atmospheric geometries. The   code is generally available to the commu­ "-   nity and uses the high resolution trans­    mission molecular absorption database HITRAN (Rothman et al., 2009) as a        molecular line database. HITRAN con­ 6@UDKDMFSGML tains energy levels, frequencies, line strengths, pressure-broadening and -shift coefficients, for more than 1.7 million   spectral lines of 42 different molecules and their common isotopes from the red TW  

optical to the sub-mm. Ck

A layered model of the Earth’s atmosphere   serves as the primary input for LBLRTM and contains temperature, ­pressure -NQL@KHRD   and molecular abundance information as a function of atmospheric height at the observatory. For a representative, yet       accessible model of the atmospheric conditions at the time of the observation,  

we supplement a static model of the "   atmosphere with meteorological data for  temperature, pressure and humidity of   the troposphere and lower stratosphere     (surface height ≤ 26 km). LBLRTM 6@UDKDMFSGML ­converts the atmospheric model into indi­ vidual isothermal layers and calculates absorption and emission spectra for a   given path through the atmosphere, i.e. for   a given zenith angle or airmass. TW Ck A single model run takes only a few sec­   onds on a standard desktop PC. The re­sulting spectrum has an intrinsic reso­   -NQL@KHRD   Figure 1. Examples of synthesised telluric spectra ­fitted to CRIRES observations of standard stars.   Each panel shows the measured spectrum in black     and the model overplotted in red. We show the residuals after division by the models, i.e., after   removal of the telluric absorption lines, at the bottom "   of each panel. Top panel: Water vapour (H2O) lines  in the H-band. Middle panel: Nitrous oxide (N2O) and   water vapour (H2O) in the L-band. Bottom panel: Ozone (O3), carbon monoxide (CO), carbon dioxide     (CO2), and water vapour (H2O) lines in the M-band. 6@UDKDMFSGML

22 The Messenger 142 – December 2010 lution that is set to resolve the narrowest Wavelength calibration as added value thus provide a natural in situ wavelength telluric lines in a given spectral region. calibration. Since they are imprinted Resolving powers are thus often of the Fitting telluric lines to empirical spectra in the source spectrum before the light order of 106 and the model spectrum not only provides a model of the telluric reaches the spectrograph, these lines needs to be convolved with the instru­ absorption, it also delivers a precise suffer the same instrumental effects as mental profile of the spectrograph to wavelength solution. This is an important the intrinsic lines in the stellar spectrum. match the measured telluric spectrum. added value of this method, given that Hence, in contrast to all emission line To first order, the instrumental profile can the wavelength calibration for high reso­ sources used to wavelength calibrate be represented by a single Gaussian lution near- and mid-infrared spectra is long-slit spectra, telluric absorption lines function. More complex profiles might be often challenging. Rare gas emission line provide an intrinsically more precise used if higher precisions and signal-to- lamps used in low resolution near-infrared calibration source, especially since lamp noise ratios need to be achieved. At the spectrographs, such as He, Ne, Xe, and emission lines are recorded in separate same time, the abundance of certain Kr lamps, provide only a sparse line den­ exposures, often hours after the science molecules might be over- or underesti­ sity and are only of very limited use at spectra and thus after potential changes mated by the atmospheric model, espe­ high spectral resolution (see also Aldenius in the spectrograph setup. cially for strongly variable species such et al., 2008). Even the rich spectra of as water vapour. Given the extreme geo­ ThAr emission line lamps, commonly em­ graphical location of most astro­nomical ployed by high resolution optical spec­ Performance and limitations observatories, including Cerro Paranal, trographs, have a much lower line density general meteorological models rarely in the near-infrared than in the red opti­ As can be seen in the examples pre­ reproduce the exact water vapour levels cal. Typical line densities of ~ 400 lines sented in Figure 1, the modelling of tellu­ on site. Thus, the model data need to per 100 nm around λ = 1000 nm drop ric absorption lines is not perfect, but be matched to the actual observing con­ quickly to less than ~ 20 lines per 100 nm leaves residuals at the 2% level for most ditions. A chi-squared fitting ­algorithm at λ = 2500 nm. Given the typical wave­ lines when observed at high spectral res­ can be employed to solve for the instru­ length coverage of ~ λ/200 per CRIRES olution. A few lines exhibit even stronger mental profile and to re-adjust the abun­ detector, many spectral settings remain residuals. The main reasons for these dance of the molecular species in the poorly calibrated. Moreover the dynamic mismatches are imprecise line data in the atmospheric model while minimising the range of the ThAr emission spectrum is HITRAN database and insufficient treat­ residuals between the ob­­served spec­ very high due to the strong contrast be­ ment of line coupling (also known as line trum and the spectral model. tween the Th and Ar lines, leaving weaker mixing) in the radiative transfer code. lines at low signal-to-noise ratios while The latter is, for example, responsible for

We show some example fits to spectra nearby stronger lines quickly saturate. the strong residuals of the CO2 Q-branch obtained with the CRIRES spectrograph at 4695 nm (see the lower panel of Fig­ in Figure 1. The resolving power was Often atmospheric OH* emission lines ure 1). Other limitations may arise from R ~ 65 000–80 000 as measured from the (commonly referred to as airglow emis­ uncertainties in the determination of the width of the instrumental profile, which sion) are used for wavelength calibration instrumental profile and the atmospheric was fitted at the same time as a single in the near-infrared, especially for low- model. Gaussian function. More examples and a and mid-resolution spectrographs. OH* detailed discussion of the fitting per­ is a chemical radical and its transitions Line data in HITRAN have strongly vary­ formance can be found in Seifahrt et al. originate from non-LTE processes in high ing accuracy levels. Typical uncertain- (2010). In all cases, the abundance of atmospheric layers. They appear in emis­ ties of line positions range from a few most atmospheric species was slightly sion but not in absorption in astronomi- to several hundred m/s, but can be as modified by a few percent during fitting to cal spectra and constitute the dominant high as several km/s in extreme cases. achieve an optimal result. The fit pro- background source in the near-infrared Line strengths are rarely precise to the cess also included the wavelength solu­ J- and H-bands. However, neither the line 1 % level. However, the HITRAN database tion for the CRIRES spectra. This step density nor the line strength of individual is constantly updated and the data qual­ is critical, since narrow and steep telluric OH* lines is high enough to make this ity will further improve in the future. Also, lines are sensitive to small wavelength species useful for the wavelength calibra­ individual line data can be re-fitted when errors and produce notable residuals for tion of high-resolution spectrographs. comparing synthesised spectra with wavelength mismatches as small as For example, the small spatial pixel scale high resolution spectra of standard stars ~ 30 m/s. A typical fit process, including of CRIRES (86 milliarcseconds/pix) makes before applying the code to spectra of various runs of LBLRTM takes about the use of OH* lines especially challeng­ science targets. two minutes to converge over the spec­ ing, given that each pixel covers less than tral coverage of a single CRIRES detec­ 10 –2 arcsecond2 on the sky. Despite these limitations, synthesised tel­ tor. We have also successfully applied luric spectra can provide the same or the code to optical high resolution spec­ In contrast, the atmospheric features of even better telluric correction than empir­ tra from UVES and to low resolution near- other regular molecules in LTE, as we ical spectra. The question of when to infrared spectra from SINFONI. model them with LBLRTM, show a high use empirical spectra or a telluric model density of strong absorption lines and depends on a number of factors and

The Messenger 142 – December 2010 23 Astronomical Science Seifahrt A. et al., Precise Modelling of Telluric Features in Astronomical Spectra

should be decided on a case-by-case in­cludes observations of early-type stars ard star observations would have com­ basis. However, the use of a telluric to map their spectra to search for unac­ promised the achievable long-term radial model should be strongly considered for counted lines. This process would prove velocity precision, strongly favouring the regions where standard stars have spec­ impossible if a standard star of similar telluric modelling approach. tral features, when instrumental setups spectral type was used as an atmospheric (including the instrumental profile) cannot calibrator. Last but not least, telluric models are be reproduced, or when observations not only used to wavelength calibrate and encompass a wide range in airmass. Another project utilising telluric models to correct for the telluric features in ob­ is the CRIRES radial velocity campaign served spectra, but also to plan observa­ Some recent and ongoing science pro­ to search for extrasolar planets around tions and predict the atmospheric grammes entirely rely on telluric models mid-to-late M-type stars (Bean et al., throughput based on meteorological fore­ for calibration. One example is the 2010a,b). The wavelength calibration for casts. For example, the exposure time CIRES–POP project, dedicated to pro­ this project relies on a custom-built am­ calculator for CRIRES employs a telluric viding the community with a high reso­ monia gas cell, given the need to deter­ model to allow the user to check spectral lution, high signal-to-noise near-infrared mine radial velocities to the m/s level. settings for the impact of telluric absorp­ spectral atlas of several stars across The telluric features present in all obser­ tion under different water vapour levels. the Hertzsprung–Russell diagram vations obtained by this programme (Lebzelter et al., 2010). Data reduction need to be corrected to a high level of for this project employs telluric models precision. Obtaining high signal-to-noise References to wavelength calibrate settings that standard star spectra for all science Aldenius, M. et al. 2008, The Messenger, 133, 14 lack coverage by ThAr lines or gas cells. observations would have made the pro­ Clough, S. A., Iacono, M. J. & Moncet, J.-L. 1992, CRIRES–POP concentrates on very ject less time-effective. Also, the instru­ J. Geophys. Res., 97, 15761 bright stars and uses a large number of mental profile between the science Rothman, L. S. et al. 2009, Journal of Quantitative CRIRES settings. Obtaining spectra observation and the standard star must Spectroscopy and Radiative Transfer, 110, 533 Seifahrt, A. et al. 2010, A&A, in press, arXiv:1008.3419 of telluric standard stars for each science match; a challenging requirement for Bean, J. L. et al. 2010a, ApJ, 713, 410 target would be very time consuming, CRIRES. Given the sensitivity of the radial Bean, J. L. et al. 2010b, The Messenger, 140, 41 at least doubling the total observing velocity method to small changes in Lebzelter, T. et al. 2010, The Messenger, 139, 33 time. Most importantly, the programme the instrumental profile, the use of stand­

NGC 3603, shown here in a VLT FORS three-colour composite (from V-, R- and I-band images), is a young Galactic starburst cluster situated at a dis­ tance of 6.8 kiloparsecs. With many dozens of mas­ sive young hot stars and an age of a few Myr, it is the focus of many studies as a nearby template for extragalactic starburst environments.

24 The Messenger 142 – December 2010 Astronomical Science

Astronomy Meets Biology: EFOSC2 and the Chirality of Life

Michael Sterzik1 molecule are called enantiomers, and the extreme environment on Earth, and the Stefano Bagnulo 2 two forms are generally referred to as microbial colonisation of subsurface lay­ Armando Azua 3 right-handed and left-handed, or dextro­ ers in halites (rock salt) and quartz rocks Fabiola Salinas 4 rotatory and levorotatory. by specific cyanobacteria. In the most Jorge Alfaro 4 hostile environments (exceptional aridity, Rafael Vicuna 3 The term homochirality is used when a salinity, and extreme temperatures), a molecule (or a crystal) may potentially primitive type of cyanobateria, Chroococ- exist in both forms, but only one is actu­ cidiopsis, can be the sole surviving 1 ESO ally present. Homochirality character- organism. This has interesting implications 2 Armagh Observatory, United Kingdom ises life as we know it: all living material for the potential habitability, and eventual 3 Department of Molecular Genetics and on Earth contains and synthesises ­sugars terraforming, of certain areas on Mars Microbiology, Pontificia Universidad and nucleic acids exclusively in their (Friedmann & Ocampo-Friedmann, 1995). Católica de Chile, Chile right-handed form, while amino acids and 4 Faculty of Physics, Pontificia Universi­ proteins occur only in their left-handed The idea of using the ESO Faint Object dad Católica de Chile, Chile representation. However, in all these Spectrograph and Camera (EFOSC2) cases, both enantiomers are chemically to investigate samples of Chroococcidi- possible and energetically equal. The opsis extracted from the underside of Homochirality, i.e., the exclusive use of ­reasons for homochirality in living mate­ Atacama Desert quartzes and to measure L-amino acids and D-sugar in biologi- rial are unknown, but they must be their circular polarisation in a laboratory cal material, induces circular polarisa- related to the origin of life. It is still dis­ experiment arose at this conference. tion in the diffuse reflectance spectra of puted whether bioactive molecules (and The idea looked appealing, because only biotic material. Polarimetry may there- with them small enantiomeric excesses) limited, and sometimes contradictory, fore become an interesting remote were delivered to Earth (e.g., by mete­or­ reports about circular polarisation meas­ sensing technique in the future search ites) or whether (pre-)biotic chemistry urements of biotic material as a signature for extraterrestrial life. We have explored started on Earth (Bailey et al., 1998). of homochirality are available in the litera­ this technique and performed a labora- Undoubtedly, however, chemical and bio­ ture. Moreover, the successful use of tory experiment making an exotic use logical processes on Earth must have an astronomical instrument for the first of an astronomical instrument. During a favoured the selection of one-handed time for that purpose could serve as a period when EFOSC2 was detached biomolecules leading towards homochi­ benchmark for further applications of this from the Nasmyth focus to host a visitor rality. If similar evolutionary scenarios method in astrophysics. instrument at the NTT, we have ob- naturally occur elsewhere in the Universe, served various samples of biotic and homochirality may be a universal hallmark A detailed feasibility study as well as abiotic material and measured their of all forms of life. the production of significant quantities of ­linear and circular polarisation spectra. Chroococcidiopsis were prepared and Among the various targets, we have in- Chirality induces optical activity: each in­itiated in the following weeks. Since cluded samples of the hypolithic cyano- enantiomer rotates the reflected (or trans­ there is no formal process in place to ob­ bacteria species Chroococcidiopsis mitted) light in opposite directions, and tain “observing time” for laboratory i­solated from the Coastal Range of the homochirality guarantees that there will experiments, the ESO Director General Atacama Desert. To our knowledge, be an excess of circularly polarised light was asked for authorisation. He approved these are the first and highest preci- in one direction. This opens up the inter­ the experiment under the condition that sion measurements of circular polarisa- esting possibility that biosignatures could it did not pose any risk to the instrument. tion using living material and obtained be sensed remotely by means of polari­ This could be achieved by using EFOSC2 with an astronomical instrument. metric techniques. when it was not attached at its nominal Nasmyth focal station (i.e. during an ex­­ This chain of arguments, which had been tended visitor instrument run), but keep­ Motivation raised in various articles in the scientific ing it in a horizontal position (which avoids literature (e.g., Wolstencroft et al., 2004), the possibility of any material falling on The building blocks of life are chiral. Their was also discussed during the workshop the entrance window). Our “laboratory” molecular structure lacks an internal Astrobio 20101 held in Santiago de Chile is shown in Figure 1, left. EFOSC2 is plane of symmetry, and their mirror image last January. For the first time, an inter­ de­tached from the New Technology Tele­ cannot be superimposed on their origi- national and interdisciplinary conference scope (NTT). A microphotograph of nal image. The term chirality is specifi­ that aimed to cover major topics in astro­ our main target, Chroococcidiopsis, is cally used when a molecule (or an object) biology was organised and hosted in displayed on the right. exists in both mirror-symmetric configu­ Chile. The topics covered included the rations. The human hands are the classic origins of life, the chemistry of the Uni­ example that illustrates the concept of verse, extrasolar planetary systems, and Experiment chirality, and the term chiral itself comes the search for life in the Solar System. from the Greek word for hand, χειρ. A prominent topic of discussion was the During one week in June 2010, three In chemistry, the two images of a chiral Atacama Desert as an example of an of us (Pontificia Universidad Católica

The Messenger 142 – December 2010 25 Astronomical Science Sterzik M. et al., Astronomy Meets Biology: EFOSC2 and the Chirality of Life

Figure 1. Left: The EFOSC2 instru- ment is shown detached from the NTT. The instrument attached to the Nasmyth focus (on the left) is ULTRA­ CAM. Right: A microphotograph of the cyanobacteria Chroococcidiopsis, enlarged 100 times.

student Fabiola Salinas, Stefano Bagnulo to measure the polarisation level with an In the Philodendron leaf we detected and Michael Sterzik) were busy using error bar of the order of 10–4 per spectral both broad polarised features (with an EFOSC2 to observe samples of minerals bin. Measurements of the cyanobacte- amplitude of ~ 0.5%), and, superposed and paints (quartz, salt, ­sugar, white ria lasted several hours, but for a total on them, a narrow feature at about ­flat-field screen), leaves (Philodendron, integration time of just about 10 minutes. 680 nm (with an amplitude of ~ 0.05% Ficus, Schefflera) and cyano­bacteria films This exposure allowed us to reach an over the continuum), both well above deposited on filter paper. All samples error bar of 10–6 per spectral bin. These the statistical noise (the green line shows were prepared as thin sheets inserted in figures refer only to the statistical errors the null profile). The behaviour of Stokes the flat-field screen position. An inte­ due to Poisson noise. With our ultra-high V in the continuum closely follows the grating sphere (the usual calibration lamp signal-to-noise ratio measurements, reflectivity, shown with a solid black line mechanism for EFOSC2 when attached we have certainly hit the limits imposed (this is known as the Cotton effect). to the 3.6-metre telescope) was used by the polarimetric optics and the ex­ This behaviour closely resembles the dif­ for diffuse illumination. EFOSC2 covers perimental conditions — we will come fuse reflectance circular dichroism large spectral regions in the spectropola­ back later to this important point. Here we spectra of leaves as seen by Wolstencroft rimetry mode: we used mostly grism note that the reliability of the error bars et al. (2004). The narrow feature around 13 to cover the range 370–930 nm, with (in terms of statistical error) has been vali­ 680 nm is very similar to the results a spectral resolution of ~ 2.3 nm (we dated with the use of so-called null- shown by Gregory & Raps (1974) in their adopted a 1-arcsecond slit width). Both profiles (i.e. the difference between Stokes transmission spectroscopy of chloro­ linear and circular polarisation measure­ profiles obtained from different pairs of plasts and is related to the chlorophyll-a ments of the samples were obtained, exposures). Null profiles were found scat­ pigment response. using the λ/2 and λ/4 retarder wave­ tered around zero within the error bars. plates, respectively (we note that the λ/4 The interpretation of the results for retarder waveplate is a recent addition to Chroococcidiopsis is the most interesting. the instrument, see Saviane et al. [2007]). Results and outlook In the continuum, Stokes V shows a behaviour similar to that observed for the Polarimetric measurements were taken Figures 2 and 3 show the results white screen field, and hence is likely by combining several pairs of expo- ob­tained from our circular polarisation to be of instrumental origin. A number of sures obtained with different position measurements (Stokes V normalised narrow, low amplitude features appear angles of the retarder waveplates (–45º, to the intensity) of a Philodendron leaf, superposed on the continuum. In order to 45º, 135º, 225º for circular polarisation, and a film of Chroococcidiopsis depos­ prove that these features are real (and and 0º, 22.5º, 45º, ... 335.5º for linear ited on filter paper, respectively. We not spurious signals, e.g., caused by the polarisation measurements, measured also obtained calibration measurements CCD readout noise) we also obtained from the principal plane of the Wollaston of a white screen flat-field, produced ultra-high signal-to-noise ratio measure­ prism). This “beam swapping” technique by a barium sulphate based white reflec­ ments with the retarder waveplate at 0º, minimises the instrumental effects — tance coating. The circular polarisation 90º, 180º and 270º. With these retarder for a detailed description of the measur­ measured for the screen flat-field was waveplate settings, we would theoreti­ ing strategies, see, e.g., Bagnulo et found as a negative continuum of about cally expect a null signal. The resulting al. (2009). For inorganic samples and for –0.05%, with the absolute value slowly profile, shown with the red line in Fig­ leaves, the typical observation cycles increasing towards shorter wavelengths. ure 3, and arbitrarily rectified to zero for lasted 20 minutes, dominated in practice No narrow features or signals are pre- display purposes, shows no high fre­ by overheads (readout time and retarder sent in the polarisation spectra of this ref­ quency signatures. This curve, together waveplate setting), and allowed us erence source. with a consistently flat null profile (shown

26 The Messenger 142 – December 2010     Figure 2. The Philoden- dron leaf is shown at left  and its polarised spec­ 

   trum measured with 5 - l  EFOSC2 at middle and right. The blue lines

l   show the circular polari­ sation (V/I). In the right­ most panel, V/I is recti­ l 

( fied to zero. The green l  5 lines show the null pro­ l  file (offset by –0.06% in the rightmost panel, l  l  for display purposes)        6@UDKDMFSGML 6@UDKDMFSGML demonstrating the sta­ tistical errors. The black line is the reflectivity (in arbtirary units).

Figure 3. The sample of     Chroococcidiopsis is shown at left and its   5 polarised spectrum in -   the middle and right   panels. The blue line  shows circular polarisa­ tion and the green line  is the null profile (offset l  ( by + 0.02%). The red 5 line shows the spectrum l  l  obtained with the re- tarder waveplate at      position angles offset by 6@UDKDMFSGML 6@UDKDMFSGML 45º, rectified to zero. In the rightmost panel, V/I is rectified to zero. with the green line offset to + 0.02% in the FORS instrument of the VLT. Some bacteria on other Solar System bodies in the right panel) suggests that the narrow of these effects may also play a role the future with extremely large telescopes features are indeed real. We notice in EFOSC2. We must also remark that to be assessed. again the prominent chlorophyll-a feature the prepared samples are not fully repro­ around 680 nm, but may also infer sig­ ducible in their configuration, and thus We thank all the La Silla staff who sup­ natures of other pigments like carotinoids likely introduce some effects due to their ported the setup and execution of this and phycocyanins. Our measurements different backscattering properties. unusual experiment! Stefano Bagnulo also appear consistent with polarisation acknowledges a grant received as visiting spectra obtained from a different type of In conclusion, the polarised spectra of scientist at ESO/Chile. cyanobacteria by Sparks et al. (2009). Figures 2 and 3 show that the polari­ metric techniques employed by us are While we kept the noise confidently at sensitive to the presence of biotic mate­ References a very low level, the polarimetric mode of rial. We detect chiral signatures of pig­ Bagnulo, S. et al. 2009, PASP, 121, 993 EFOSC2 is lacking the full characterisa­ ments involved in biotic photosynthetic Bailey, J. et al. 1998, Science, 281, 5377, 672 tion needed to reach ultra-high and abso­ reaction chains. Signatures of chlorophyll Bailey, J., Lucas, P. W. & Hough, J. H. 2010, lute precision polarimetric measure- pigments can be captured both in sam­ MNRAS, 405, 2570 ments under the conditions of our experi­ ples of leaves and cyanobacteria of spe­ Gregory, R. P. F. & Raps, S. 1974, Biochem. J., 142, 193 ment. Using special calibration tech­ cies Chroococcidiopsis. More accurate Friedmann, E. I. & Ocampo-Friedmann, R. 1995, niques, in some astronomical observa­ measurements of the polarised signal Adv. Space Res., 15(3), 243 tions it is already possible to achieve a amplitudes require experiments to be Patat, F. & Romaniello, M. 2006, PASP, 118, 146 precision of 10–6 (e.g., Bailey et al., 2010), carried out with samples prepared under Saviane, I. et al. 2007, The Messenger, 129, 14 Sparks, W. B. et al. 2009, PNAS, 106 (19), 7816 but in our experiments we are limited better controlled laboratory conditions. Wolstencroft, R. D., Tranter, G. E. & Le Peleven, D. D. by an instrument setup in which most of It will also be interesting to observe the 2004, Bioastronomy 2002, IAU Symp., 213, 149 the signal comes from off-axis rays. Patat polarimetric signatures of Chroococcidi- & Romaniello (2006) and Bagnulo et al. opsis in their natural habitat beneath (2009) have discussed the spurious polar­ translucent rock surfaces. This will allow Links imetric effects observed off-axis with the feasibility of detecting this type of 1 http://www.astro.puc.cl/astrobio2010

The Messenger 142 – December 2010 27 Astronomical Science

Observations of Multiple Stellar Populations in Globular Clusters with FLAMES at the VLT

Raffaele Gratton1 Eugenio Carretta2 Angela Bragaglia2 Sara Lucatello1 Valentina D’Orazi1

1 INAF–Osservatorio Astronomico di Padova, Italy 2 INAF–Osservatorio Astronomico di Bologna, Italy

In the last few years, it has become evi- dent that globular clusters, previously accepted as prime examples of simple stellar populations, contain at least two, and in some cases more, stellar generations. Thanks to the superb spa- tial resolution of the Hubble Space ­Telescope on one side, and to the fore- front spectroscopic and multiplexing capabilities of UVES and FLAMES at the VLT on the other, extensive, high preci- sion datasets are shedding light on clus- ter formation and evolution. Here, we briefly describe the contribution by our team to these exciting discoveries. Figure 1. Image of the globular cluster 47 Tuc This abundance pattern between O–Na (NGC 104), obtained with the ESO 1-metre Schmidt and Al–Mg is characteristic of GCs, Telescope at the La Silla Observatory in Chile. while it is almost absent among field stars Multiple populations in globular clusters (Gratton et al., 2000). It is not limited All GCs investigated to date show a pecu­ to red giants, but is present also in the Globular clusters (GCs) are very massive liar abundance pattern (for a review, see low-mass MS stars, as demonstrated stellar complexes. As an example, Fig­ Gratton et al., 2004). They usually appear by the key study with the VLT’s high reso­ ure 1 shows an image of the globular clus­ to be very homogeneous, with a few lution optical spectrograph, UVES, by ter 47 Tucanae. All GCs observable in exceptions like M 54 and Omega Centauri, Gratton et al. (2001; based on the large or near the Milky Way are very old, so their insofar as the Fe-peak elements are con­ programme 165.L-0263). These un­- formation cannot be observed in detail. cerned, thus ruling out self-enrichment by evolved stars cannot have sustained the These early phases were likely complex, SN ejecta. However, many GC stars have nucleosynthesis chains that deplete O including a variety of dramatic and ener­ an unusual composition, rich in Na and and Mg, and enhance Na and Al, because getic phenomena such as supernova Al and poor in O and Mg, typical of mate­ they cannot reach the required tempe­ (SN) explosions, photoionisation, high and rial that experienced H-burning at very ratures and have very thin convective en­ low velocity winds from blue and red high temperatures, while others have a velopes, unable to mix nuclear prod- massive stars, possibly combining in giant normal O- and Mg-rich but Na- and Al- ucts into their atmospheres. The UVES expanding bubbles of gas and shock poor composition, which is virtually iden­ large programme excluded early claims fronts that may have triggered further star tical to that of metal-poor field stars. that this composition might result from a formation. The scene where this drama Within each GC, the abundances of Na peculiar evolution of GC stars: all low occurred might have been even more and O, or Al and Mg are anti-correlated O/high Na stars originated from matter di­­versified — in the core of giant clouds with each other (see Figure 2). Since processed and ejected by stars belong­ or, for the most massive clusters, even the main outcome of H-burning is He, we ing to a previous stellar generation within of dwarf galaxies. These dramatic events expect low O/high Na stars also to be the GC, although not by the SNe (or left a trace — represented by the chemi­ He-enriched, which explains the occur­ the low-mass stars, which give a different cal composition of the stars — that not rence of multiple HBs and MSs. The characteristic imprinting) of this earlier only can be followed today in quite subtle ­connection between O–Na and Mg–Al population. spectral features, but also can influence anti-correlations, variations in He abun­ their distribution along the main se­quence dances, and the presence of multiple Self-pollution models still lag behind the (MS) and the horizontal branch (HB). HBs is however likely to be more complex observations and are not yet able to than described here (see Gratton et al., ­explain convincingly all observed features: 2010). candidate first generation (FG) polluters

28 The Messenger 142 – December 2010 are either thermally pulsating intermediate-  mass asymptotic giant branch (AGB) stars undergoing hot bottom burning   (Ventura et al., 2001) or massive rotating

-@%D  stars prior to SN explosion (Decressin et -&" -&" -&" -&" -&" al., 2007). l  

Our FLAMES survey  

-@%D  Although the star-to-star O–Na abun­ -&" -&" -&" -&" -&" dance anticorrelation had been recog­ l  nised early on as a pivotal signal of  ­self-enrichment in GCs, significant pro­ gress was hindered by the slow acqui­   sition rate of data, due to the use of

-@%D  ­single object spectrographs. At the time -&" -&" -&" -&" -&" of the Gratton et al. (2004) review, data l  for a grand total of only about 200 stars,  distributed in some ten GCs, were availa­ ble. The high multiplexing capability of   the VLT’s Fibre Large Array Multi Element

-@%D  Spectrograph (FLAMES) allowed us -&" -&" -&" -&" -&" an order of magnitude increase in data l  collection capability: our survey has l l l l l  already harvested spectra for more than %D %D %D %D %D 2000 red giants, analysed individually with a homogeneous procedure. With our Figure 2. The Na–O anticorrelation is shown as least two thirds of the total population. work, we intended to answer some fun­ log(Na/Fe) vs. log(O/Fe) normalised to the solar Since the peculiar chemical compo- abundance in 20 GCs as observed with FLAMES damental questions: Were GC stars really (adapted from Carretta et al. [2009a] and Carretta sition of the SG stars indicates that only born in a single instantaneous burst? et al. [2010]). The typical error bar for each set of a fraction of FG stars may have pro- Did all GCs self-enrich themselves? How measurements is shown. duced the right gas mixture, the original do abundance patterns within each clouds from which the GCs formed ­individual GC relate to the formation and should have been much more massive early evolution of the GC itself and of each anticorrelation”, as distinct from associa­ than current GCs (likely ten times or even individual member? tions and open clusters (Carretta et al., more). Since GCs now account for 2010a). In Figure 3, adapted from Carretta roughly 1% of the Galactic halo, a large We selected a large sample of GCs with et al. (2010a), we show massive stellar fraction of the halo should have origi­ diverse HB morphology, since, as men­ clusters (both GCs and open clusters) in nated in the same episodes that led to tioned, there appears to be a link between the relative age vs. mass plane. Here the formation of the GCs. the chemical anomalies, the extension ­relative ages are used (see Carretta et al., of the HB, and the presence of multiple 2010a): in this scale, a relative age of We obtained diverse distributions of the MSs. We obtained FLAMES data on 1 means an age close to 12.5 Gyr. Differ­ Na–O and Mg–Al anticorrelations for 24 GCs (ESO programmes 072.D-0507, ent symbols are used for open clusters ­different GCs (Carretta et al., 2009a, Fig­ 073.D-0211, 081.D-0286, and 083.D-208). and GCs, and for those clusters where ure 2; Carretta et al., 2009b) in some We used the pipeline-reduced spectra the Na–O anticorrelation has been found of the GCs, indicating that some parame­ to obtain Fe, Na, and O abundances from or not (for several clusters, data currently ter, most likely the typical polluter mass, spectra from the medium-high resolution available are not sufficient to clarify this is varying from cluster to cluster. There is spectrograph, GIRAFFE, and of more point). From this figure, it seems clear a strong link between the extension of elements, including Mg and Al, from the that the presence of multiple populations the Na–O anticorrelation and some of the UVES data. Up to now we have published is the typical result of the formation pro­ main parameters of GCs, mainly mass about 15 refereed papers presenting re­ cess of massive stellar clusters and must and metallicity. sults for 21 GCs. be explained by their formation scenarios. We also confirmed that in spite of these With this data, we have confirmed that The size of our sample led, for the first striking star-to-star differences in the all GCs contain multiple populations time, to a quantitative estimate of the light elements, most GCs (with a few ­(Carretta et al., 2009a). We even pro­ fraction of FG and second generation notable exceptions) are extraordinarily posed a new definition of bona fide GCs (SG) stars in GCs (Carretta et al., 2009a). homogeneous in Fe-peak elements, with as “stellar aggregates showing the Na–O The SG is always dominant, including at upper limits as low as 5% on the root

The Messenger 142 – December 2010 29 Astronomical Science Gratton R. et al., Observations of Multiple Stellar Populations in Globular Clusters

Figure 3. The relative age parameter is plotted vs. absolute magnitude MV for globular and old open clusters. Red filled pentagons are for GCs where the Na–O anticorrelation has been observed. Open stars l mark GCs for which insufficient data about the Na–O  — ,ी anticorrelation is available. Squares are for the clus­ ters of the Sagittarius dSph galaxy and open circles are for old open clusters (see Carretta et al., 2010a). Green crosses mark those clusters where the Na–O l anticorrelation has been searched for but not found. Superimposed are lines of constant mass (light solid — , lines, see Bellazzini et al., 2008). The heavy blue ी 4 solid line (at a mass of 4 × 10 MA) is the proposed separation between globular and open clusters. l 5 , mean square dispersion. A by-product of our work is a new metallicity scale for — , GCs based on homogeneous abun­ l ी dances from high resolution UVES spec­ tra (Carretta et al., 2009c). The homo­ geneity and good statistics also allowed the first distinction between the He-rich, l — , SG and the He-poor FG populations ी at the red giant branch (RGB)-bump ­(Bragaglia et al., 2010a) to be made, as foreseen by models.        1DK@SHUD@FD Our first extensive study of the chemical composition of M 54 in the nucleus of environments: our finding is then a direct HB, or the blue and red sequences of the disrupting Sagittarius dwarf galaxy probe of the ambient condition at the dis­ GCs seen in many galaxies), and the showed a metallicity dispersion and Na–O tant epochs where the bulk of different ­relation between GCs and the nuclei of anticorrelation in both the metal-rich stellar generations formed in GCs. dwarf galaxies, and many more. and metal-poor component (Carretta et al., 2010b), with similarities with Omega These results are unprecedented and With its large set of very competitive in­ Centauri, probably indicative of a similar were made possible by the large samples struments, the VLT is likely to play a origin in dwarf spheroidal galaxies. of high resolution spectra available, ­fundamental role in this game even in the and by the high degree of homogeneity future. UVES and FLAMES observa- The Li abundances offered a comple­ attained by the analysis procedures tions were essential in revolutionising our mentary approach. Li is easily destroyed adopted in our survey. view of the GCs. New perspectives are in stellar interiors so, if there is no Li possible with the high sensitivity of ­production within the polluters, Li and O X-shooter, revealing the spectra of faint should be positively correlated and Li Future directions MS stars (see Bragaglia et al., 2010b), or and Na anticorrelated. Measures of O, Na, the access to near-infrared spectra pro­ and Li abundances in the same stars Our results, as well as the discovery of vided by CRIRES. are rare, but D’Orazi and Marino (2010) multiple MSs from high precision found no Li–O correlation in a sample Hubble Space Telescope photometry, are of about 100 giants in M4. If confirmed in opening unexpected and fascinating References other GCs, this would be explained ­windows on the quest for the formation Bellazzini, A. et al. 2008, MSAIt, 79, 663 only with Li production in the polluters, and evolution of massive stellar clus- Bragaglia, A. et al. 2010a, A&A, 519, 60 very likely intermediate-mass AGB stars. ters. However, several issues remain Bragaglia, A. et al. 2010b, ApJL, 720, L41 presently unsolved, like the nature of the Carretta, E. et al. 2009a, A&A, 505, 117 Finally, we found that the fraction of polluters, the overall timescale of the Carretta, E. et al. 2009b, A&A, 505, 139 Carretta, E. et al. 2009c, A&A, 508, 695 Ba-stars (which arise from mass transfer ­self-enrichment, the existence of material Carretta, E. et al. 2010a, A&A, 516, 55 in binaries) is higher among FG stars with pristine composition diluting the Carretta, E. et al. 2010b, ApJL, 714, L7 (D’Orazi et al., 2010). This prompted us to products of the polluters, the relation Decressin, T. et al. 2007, A&A, 464, 1029 estimate the binary fraction from multi- between the formation of GCs and that of D’Orazi, V. & Marino, A. F. 2010, ApJL, 716, L166 D’Orazi, V. et al. 2010, ApJL, 719, L213 epoch measurements of radial velocities, the Galactic Halo, the role of different Gratton, R. et al. 2000, A&A, 354, 169 available for three of our GCs, finding that populations on the dynamical evolution of Gratton, R. et al. 2001, A&A, 369, 87 the binary fraction of SG stars is much GCs, the connection between multiple Gratton, R. et al. 2004, ARA&A, 42, 385 lower than that of FG stars. Binaries are populations and other properties of GCs Gratton, R. et al. 2010, A&A, 517, 81 Lata, S. et al. 2002, A&A, 388, 158 more common in low than in high density (e.g., mass, the second parameter on the Ventura, P. et al. 2001, ApJ, 550, L65

30 The Messenger 142 – December 2010 Astronomical Science

Dissecting the Galactic Super Star Cluster Westerlund 1 — A Laboratory for Stellar Evolution

Simon Clark1 Ignacio Negueruela 2 Ben Ritchie1 Paul Crowther 3 Sean Dougherty4

1 Department of Physics and Astronomy, The Open University, Milton Keynes, United Kingdom 2 Departamento de Física & Ingeniería, Universidad de Alicante, Spain 3 Department of Physics and Astronomy, University of Sheffield, United Kingdom 4 National Research Council, Herzberg Institute for Astrophysics, Penticton, Canada

Westerlund 1 is the first example of a super star cluster identified within the Galaxy. As such, its proximity allows us to resolve directly and study individ- ual stars down to sub-solar masses, as well as their complex interactions. This in turn permits advances in our un­derstanding of the physics of these stellar powerhouses, which drive evo­ lution in starburst galaxies near and far. Here we provide a brief overview of our current understanding of this clus- ter, both in terms of its stellar constit­ Figure 1. The central region of the vigorous starburst uents — and the constraints they place galaxy NGC 1313 is shown in this colour image from VLT FORS data (from Release eso0643). Note on the evolution of massive stars from the numerous shells of ionised gas excited by young cradle to grave — and its global proper- massive clusters similar to Wd1. ties.

Moreover, the same population of mas­ within such clusters follow comparable Starburst galaxies and super star clusters sive stars provides significant input evolutionary paths to those within of mechanical energy and chemically the Local Group, such that we under­ Images of starburst galaxies such as processed material via their stellar winds stand their radiative output, lifetimes and the Antennae, M82 and NGC 1313 (see and SNe. Indeed, the combination of post-SNe endpoints. Without such Figure 1) reveal that the process of star both radiative and kinetic feedback is assumptions we cannot accurately cali­ formation appears hierarchical, with stars thought to be responsible for the initiation brate the population synthesis codes forming within massive clusters, which of superwinds that may promote or sup­ used to determine cluster ages and inte­ in turn are located within larger complexes press subsequent generations of star grated masses (and hence star formation that reflect the underlying structure of ­formation, as well as enrich the inter­ rates), nor constrain the degree of the natal giant molecular clouds. The im­ galactic medium. Nevertheless, despite mechanical and radiative feedback that pact of both these super star clusters their pivotal role in galactic evolution the yields galactic-scale superwinds. (SSCs) and complexes on the wider galac­ small physical extent of SSCs in exter- tic ecology is immense: the OB stellar nal galaxies means that they must be Unfortunately, through the 20th century population contained within dominates studied via comparison of their integrated it had been supposed that the Gal- both the ultraviolet (UV) and, via dust re­­ spectral and photometric properties. axy lacked spatially-resolved examples of processing, the infrared (IR) radiative SSCs, with the integrated masses of output of the host galaxy, while their post- But such an approach is based on two young (< 20 Myr) massive open clusters 3 supernovae (post-SNe) relativistic rem­ unproven hypotheses: (i) that the Initial typically being less than 10 MA com­ 5 7 nants — neutron stars and black holes Mass Function (IMF) of stars within such pared to 10 –10 MA for the former. Tar­ within binary systems — are responsible clusters (and complexes) is identical to geted near-IR observations of the centre for the high energy emission. that determined locally; and (ii) that stars of the Galaxy had revealed the presence

The Messenger 142 – December 2010 31 Astronomical Science Clark S. et al., Dissecting the Galactic Super Star Cluster Westerlund 1



 ¦o 



  A

 A  )   @  @ 

@S HN M  lo A A #DBKHM 

lo

  

 GLR L R LR LR 1HFGS RBDMRHNM)

Figure 2. Optical image of Wd1 obtained with Figure 3. Overplot of radio observations (contours) the Wide Field Imager mounted on the MPG/ESO on an optical image of Wd1. Note the emission 2.2-metre telescope. Note the high degree of associated with the cool hypergiants (Wd1-4, 12a, reddening to the cluster in comparison to the fore­ 16a, 32 & 265 — YHGs; Wd1-20, 26 & 237 — RSGs) ground B supergiant (lower right). Image from indicative of significant ongoing mass loss. Release eso1034c. of the Arches, Quintuplet and Galactic the B[e] supergiant Wd1-9 serendipi- the presence of low excitation metallic 4 Centre clusters with masses ~ 10 MA; tously detected a large number of radio and molecular features for cool objects, still an order of magnitude smaller than sources amongst the evolved stellar was constructed from both real and known SSCs. In fact the first known ­population (Figure 3; see Clark et al., ­synthetic spectra (Clark et al., 2005; example of a SSC had been identified 1998 and Dougherty et al., 2010). Unex­ Negueruela et al., 2010). As well as con­ fully 40 years earlier — indeed before pectedly, a number of the cool red super­ firming the classifications of Westerlund their widespread recognition in external giants (RSGs) and yellow hypergiants (1987), these observations also identi- starburst galaxies! — but had subse­ (YHGs) were found to be strong radio fied large additional populations of OB quently remained in relative obscurity. sources, despite lacking the requisite UV supergiants and Wolf–Rayet stars (WRs), flux to ionise their ejecta. which had previously escaped detection due to a combination of comparatively Westerlund 1 Prompted by these results we undertook low optical luminosity and heavy redden­ spectroscopic and photometric ob­ ing. Indeed, subsequent observations Located within the constellation of Ara servations of cluster members between revealed a ­population of 25 WRs, com­ (the Altar) Westerlund 1 (Wd1; see ­2001–2; firstly with the ESO 1.52-metre posed of both WN and WC subtypes; the ­Figure 2) was simply described by Bengt telescope and the Boller & Chivens spec­ richest haul detected within a Galactic Westerlund in 1961 as a very young, trograph, and subsequently at higher cluster at this time (Crowther et al., 2006). “heavily reddened cluster”. Indeed the ­signal-to-noise ratio (S/N) and resolution high extinction towards Wd1 (Av ~ 11 mag) with the New Technology Telescope These observations enabled us to deter­ hampered spectroscopic investigation, and EMMI spectrograph (Clark et al., mine that Wd1 has an age of ~ 4–5 Myr with the first such survey following over a 2005). Given the reddening towards Wd1, and appears to be co-eval; conclusions quarter of a century later (Westerlund, classification spectra were obtained confirmed by later observations (e.g., 1987). Despite revealing an unprece­ from 600–900 nm, rather than the more Negueruela et al., 2010). Moreover, utilis­ dented population of high luminosity common 400–600 nm window. An ing these spectra to calibrate our pho­ supergiants of both early and late spec­ appropriate classification scheme, based tometry, we were able to estimate that tral types, Wd1 once again sank back on the occurrence and strength of H i, >> 100 stars within the cluster must into obscurity until radio observations of He i and He ii lines in the OBA stars and have evolved from ­progenitors with initial

32 The Messenger 142 – December 2010     Figure 4. Comparison of masses ≥ 30 MA. Such a census has observed (black) and two implications. Firstly, it offers an - ( synthetic (red) spectra of explanation for the anomalous radio emis­   the LBV Wd1-243, sion associated with the YHGs and focusing on the N i and #(! RSGs, whereby the diffuse UV radiation     O i triplets and indicat- field from the hot stellar population ing significant chemical

TW ­evolution of N and O (OB stars and WRs) ­provides the ionising    

Ck (respectively 12 and 0.1 photons. Secondly, using this popula- times solar abundances). tion to normalise a Kroupa-type IMF we     were able to infer a total (initial) mass for 5 -NQL@KHRD Wd1 of ~ 10 MA, making it the most     massive Galactic cluster yet discovered by an order of mag­nitude and hence the     first SSC within the Milky Way.

    . ( Therefore, Wd1 is the first example of an              SSC for which its relative proximity per­ 6@UDKDMFSGML 6@UDKDMFSGML mits the resolution and study of individual stars to sub-solar masses — enabling the ecology of such an agglomeration to tionary states that include the luminous with initial masses in excess of ~ 25 MA be decoded for the first time. blue variables (LBVs) and the closely give rise to the post-SNe formation of related P Cygni supergiant/B hypergiant, black holes rather than neutron stars. supergiant B[e] stars, YHG and RSG Consequently, with a current MS turnoff

Massive stellar evolution — a tale of two stages. It is thought that mass loss in mass in excess of ~ 30 MA (Clark et pathways such phases may play a dominant role in al., 2005; Negueruela et al., 2010), it had the removal of hydrogen and subse- been expected that only black holes A key driver in massive stellar evolution quent formation of WRs, but it appears should currently be forming within Wd1. is mass loss, which strips away the H-rich likely that this process is accomplished Indeed, given that the pre-SNe mass mantle of O stars to yield H-depleted via transient outbursts in which mass- loss required to yield a neutron star rather WRs. Indeed the rate at which this pro­ loss rates increase by several orders of than a black hole is greatly in excess ceeds not only governs the precise evo­ magnitude over their quiescent values. of that expected for single star evolution, lutionary path trodden by the star, but A prime example of this behaviour in significant binary driven mass loss also its ultimate post-SNe fate. Radio ob­­ LBVs is the 19th-century eruption of appears mandatory to allow for the for­ servations had already revealed the h Car, while similar behaviour has also mation of the Wd1 magnetar. ­characteristic signature of heavy mass been inferred for YHGs and RSGs such loss associated with many stars (Clark et as ρ Cas and VY CMa. However, such al., 1998; Dougherty et al., 2010), while outbursts are rare, with none subject to Quantitative analysis via VLT follow-up our new spectroscopy reveals a rich post- modern day observational and analytical main sequence (post-MS) population. techniques, meaning that the physical The existing spectral data were insuffi­ Thus Wd 1 presents a unique opportunity mechanism leading to the extreme mass- cient to address these issues, being of to investigate the physics of stellar evo­ loss events has yet to be identified. too low S/N and resolution to enable lution for some of the largest stars to be non-LTE-model atmosphere analysis and found within the Galaxy. Finally, a synthesis of our multi-wavelength only at a single epoch, thus preventing observations of Wd1 suggests a very the identification of either intrinsic (insta­ Nevertheless, such a goal is challenging high binary fraction amongst the WR pop­ bilities) or extrinsic (binary) variability. for three reasons. Firstly, sophisticated ulation (Crowther et al., 2006; Clark Consequently, throughout the period modelling with non-local thermodynamic et al., 2008). Binary-mediated mass loss 2004–9 we obtained multiple epochs of equilibrium (non-LTE) atmospheric codes through Roche-lobe overflow in short high quality, 600–900 nm spectroscopic is required to determine accurately physi­ period systems offers an additional evolu­ data of ~ 100 members of the evolved cal properties such as stellar luminosity, tionary pathway for massive stars, stellar population of Wd1 with both FORS ­temperature, mass-loss rate and chemi­ which leads to significantly lower pre-SNe and FLAMES on the VLT. Full details of cal abundances; all of which are essential core masses than expected for single the target selection, experimental setup to determine accurately the precise evo­ stars. The importance of this pathway was and reduction techniques employed may lutionary state of an individual star (see dramatically demonstrated by the dis­ be found in Ritchie et al. (2009a) and Figure 4; Ritchie et al., 2009b). Secondly, covery of a magnetar (a class of neutron Negueruela et al. (2010). These data ena­ as such stars transit from the MS to star with magnetic field strengths bled us to address each of the challenges H-depleted WR phase they pass through ~ 1015 G) within Wd1 (Muno et al., 2006). described above. a dizzying variety of short-lived evolu­ Conventional wisdom assumes that stars

The Messenger 142 – December 2010 33 Astronomical Science Clark S. et al., Dissecting the Galactic Super Star Cluster Westerlund 1

Firstly, the data permit quantitative atmos­ Figure 5. Time-resolved 2H ( 2 ( %D ( %D ( ,F ( ,F ( 2H ( pheric modelling of individual stars  K@ spectra of the pulsating /@ /@ YHG Wd1-265. Compar­ in order to determine their stellar parame­ ison with classification ters. An example of such an analysis  spectra (dotted lines) is presented in Ritchie et al. (2009b; see indicates significant vari­   Figure 4). By constraining the chemical ability in spectral type %K@ over a period of only 46 abundances for the LBV Wd1-243 it days. was possible to demonstrate unambigu­  ously that it is a highly evolved, likely  post-RSG, object. The use of elemental abundances to place the panoply of %K@A post-MS stars — LBVs, B/YHGs, RSGs,

-NQL@KHRDCkTW NEERDS   etc. — in a precise evolutionary scheme is a powerful technique, with mass loss systematically driving down the H/He -( ratio while simultaneously exposing the  products of nuclear burning at the stel-       lar surface. We aim to extend this ap­ 6@UDKDMFSG proach to the remaining population of transitional stars within Wd1 in the imme­  Figure 6. Radial velocity diate future. curve for both compo­ nents of the WNL+OB binary Wd1-13 folded on Secondly, in combination with existing  the ~ 9.27-day orbital historical observations, the data ena- period. bled a search for variability amongst the subset of bright evolved stars over a  ~ 50-yr baseline. Such an approach had previously allowed for the classification of Wd1-243 as an LBV, while radio obser­ l vations of the B[e] star Wd1-9 showed an episode of enhanced mass loss that 1@CH@KUDKNBHSXJLR likely ended within the last 200 years l (Dougherty et al., 2010). Intriguingly, while ARNQOSHNM significant wind variability and pulsa- $LHRRHNM tional instability appeared to be ubiqui­ l tous for all subtypes of evolved stars l           /G@RD observed within Wd1, no further exam­ ples of LBVs have yet been identified within the hot super-/hyper-giant popula­ modelling, continued spectral monitoring (FLAMES) — the identification of radial tion. of these stars raises the prospect of velocity variables resulting from binary both determining the duty cycle of such motion. While analysis of the full dataset However dramatic variations in spectral mass-loss events as well as providing is currently underway, preliminary re- type, likely reflecting pulsation-driven stringent constraints on the underlying sults for the first year of observations of changes in the stellar photosphere, were physics driving these instabilities. Like­ the brightest quarter of targets imply a identified amongst the Blue-/YHG and wise, high cadence observations of binary frequency of > 40 % amongst the RSG populations (c.f. the YHG Wd1-265 the hotter supergiants potentially open up OB-supergiant population (Ritchie et al., shown in Figure 5; Clark et al., 2010); their internal structure to scrutiny via 2009a). This finding suggests that single indeed all stars of spectral types later asteroseismology studies, while also prob­ and binary channels may be of compa­ than B1 appear to be pulsationally unsta­ ing the origin of their highly structured rable importance in the evolution of mas­ ble. Amongst the cool super-/hyper- stellar winds. sive stars, although a determination of giants this occurrence is of considerable the period distribution of the binary popu­ importance since identical behaviour Finally, we turn to the main science lation will be necessary to quantitatively has been observed in field YHGs such as driver for our intensive spectrographic confirm this assertion. ρ Cas, where it has been associated survey — comprising multi-epoch obser­ with episodes of dramatically enhanced vations of ~ 100 evolved stars over a A second critical result from this analysis –2 –1 mass loss (~ 5 × 10 MA yr ; Lobel et 14-month baseline utilising the VLT Fibre was the identification of Wd1-13 — which al., 2003). When coupled with quantitative Large Array Multi Element Spectrograph had previously been identified as an

34 The Messenger 142 – December 2010 eclipsing system — as a double-lined the binary population of Wd1 in terms of this process? Answers to both ques­ WNL+OB supergiant binary (Figure 6). of frequency, orbital separation and mass tions are essential if we are to understand This combination of properties enabled ratio, which will be accomplished by the impact of SSCs on their wider (extra-) us to determine dynamical masses of comparison of the complete RV dataset galactic environments.

23 ± 3 and 35 ± 5 MA for the WNL and to Monte-Carlo simulations. It is hard OB supergiant components respectively to overestimate the importance of such a And finally regarding Wd1 in a wider con­ (Ritchie et al., 2010). Comparison to theo­ goal — in addition to constraining the text — did it form alone? Examination retical models of WR binary evolution relative weighting of single and binary evo­ of star-forming regions suggests that such suggest that the WNL component had an lutionary channels accurately, such infor­ clusters do not form in isolation, but initial mass of ~ 40 MA, which immedi­ mation will also constrain the physics ­currently there is no evidence for ongoing ately places a firm lower limit on the mass of massive star formation (e.g., Ritchie et star formation closely associated with of the magnetar progenitor. Given that al. [2009a] and references therein) as Wd1. Was it born with other siblings which magnetars are also present in clusters well as the production efficiency of both have since dispersed, or did it instead with comparatively low-mass MS turnoffs high- and low-mass X-ray binaries, bi- form monolithically in a single starburst

(< 20 MA; Clark et al., 2008), their pro­ nary pulsars — a major source of gravita­ event — and if so, why? How many other genitors clearly span a range of masses, tional radiation — and potentially the pro­ examples lurk in the Galactic plane? implying that an additional “ingredient” duction of gamma-ray bursts. ­Systematic spectroscopic follow-up of such as rapid rotation or a high natal candidates identified via current and magnetic field must be required for their Looking further ahead, the presence of future surveys such as the VISTA Varia­ formation. RSGs within Wd1 permits a clear and un­- bles in the Via Lactea survey will enable equivocal test of current theories of mas­ us to place and understand Wd1 in the Moreover, we may also place quantitative sive stellar evolution, which do not permit context of the recent star formation his­ constraints on the location of the bifurca­ such stars to exceed log(L/LA) ≥ 5.6 tory of the Milky Way. Clearly the discov­ tion in the canonical “Conti scenario” for (e.g., Meynet & Maeder, 2005). Unlike ery and analysis of Westerlund 1 has stellar evolution, whereby the most mas­ field stars, the well-determined distance answered many questions, but has raised sive stars evolve via: to Wd1 permits an accurate determina­ many more. tion of the luminosity of the cluster RSGs O MS → WNLh → LBV → WNL → WNE once an appropriate bolometric cor­ → WC → SNe rection has been determined via model References atmosphere analysis. Clark, J. et al. 1998, MNRAS, 299, L43 thus avoiding a cool hypergiant phase, Clark, J. et al. 2005, A&A, 434, 949 and less massive stars via Turning to Wd1 as a whole, we have a Clark, J. et al. 2008, A&A, 477, 147 unique opportunity to investigate the in­ Clark, J. et al. 2010, A&A, 514, 87 O MS OB SG RSG LBV/YHG? teraction between individual massive Crowther, P. et al. 2006, MNRAS, 372, 1407 → → → → Dougherty, S. et al. 2010, A&A, 509, 79 WNL → WNE(→WC?) → SNe stars within an SSC. The motivation for Lobel, A. et al. 2003, ApJ, 583, 923 such a study is vividly illustrated by the Meynet, G. & Maeder, M. 2005, A&A, 429, 581 (where the WNLh designation means that cometary nebulae associated with the Muno, M. et al. 2006, ApJ, 636, L41 the star shows hydrogen emission lines RSGs Wd1-20 & 26 and which are visible Negueruela, I., Clark, J. & Ritchie, B. 2010, A&A, 516, A78 in its spectrum). Given the both YHGs and in mid-IR and radio images (Figure 3; Ritchie, B. et al. 2009a, A&A, 507, 1585 RSGs are currently present within Wd1, Dougherty et al., 2010). These appear to Ritchie, B. et al. 2009b, A&A, 507, 1597 the masses implied for the progenitors of be the result of the ablation and en­ Ritchie, B. et al. 2010, A&A, 520, 48 these stars by Wd1-13 reveals that the trainment of the outer stellar layers/winds Westerlund, B. 1987, A&AS, 70, 311 division must occur at > 40 MA. of these stars by the incipient cluster wind driven by the mechanical and radia­ tive feedback from individual stars Future prospects and SNe. But how do such winds work? Comparison of radio, mid-IR and X-ray In the last decade of study Wd1 has data indicate that the intercluster medium yielded many of its secrets, enabling us appears to be multi-phase, composed to confirm that it is the first SSC to be of cool, neutral and dust-laden clumps identified within the Galaxy, as well as shadowing warmer ionised gas in close permitting us to place powerful observa­ proximity to the core, which, in turn, tional constraints on the evolution of are imbedded in an X-ray-bright compo­ massive stars in their natural environment nent, emitting via both thermal and non- for the first time. However, what does thermal mechanisms. How is momen- the future hold? Our immediate goal is to tum imparted to this material to yield a fully determine the physical properties of cluster wind and what is the efficiency

The Messenger 142 – December 2010 35 Astronomical Science

AMAZE and LSD: Metallicity and Dynamical Evolution of Galaxies in the Early Universe

Roberto Maiolino1 being more metal-poor. In some of mental relation between metallicity, mass Filippo Mannucci 2 these massive galaxies we can even and SFR observed in local galaxies. This Giovanni Cresci 2 map the gas metallicity. We find that finding suggests that the same mecha­ Alessio Gnerucci3 galaxies at z ~ 3.3 have regular rota- nisms of galaxy formation are at work Paulina Troncoso1 tion, though highly turbulent, and at any epoch in the redshift interval Alessandro Marconi 3 inverted abundance gradients relative 0 < z < 2.5. Francesco Calura4 to local galaxies, with lower abun- Andrea Cimatti 5 dances near the centre, close to the Until recently, little was known about the Filomena Cocchia1 most active regions of star formation. metallicity of galaxies at z > 3, due to Adriano Fontana1 Overall the results suggest that promi- the difficulty of detecting the optical neb­ Gianluigi Granato 6 nent inflow of pristine gas is respon- ular lines (redshifted into the near-infra­ Andrea Grazian1 sible for the strong chemical evolution red) required to measure the metallicity in Francesca Matteucci7 observed in galaxies at z > 3. these faint systems. However, this is Taro Nagao 8 a crucial epoch of very fast galaxy evolu­ Laura Pentericci1 tion, just before the peak of cosmic star Antonio Pipino 9 The chemical enrichment that we ob­ formation, which requires thorough inves­ Lucia Pozzetti10 serve in local galaxies has been pro­ tigation to understand the formation of Guido Risaliti 2 duced by the nucleosynthesis of stars primeval galaxies properly. Laura Silva 6 formed over their cosmic lives. Such enrichment has been modulated by the inflow of ­pristine gas (which both boosts The AMAZE and LSD surveys 1 INAF–Osservatorio Astronomico di star formation and dilutes the gas metal­ Roma, Monte Porzio Catone, Italy licity), enriched gas outflows (driven by We have undertaken two major projects 2 INAF–Osservatorio Astrofisico di Arce­ the star formation activity itself or by using the near-infrared integral field tri, Firenze, Italy active galactic nuclei [AGNs]) and gas ­spectrograph SINFONI at the VLT. AMAZE 3 Dipartimento di Fisica e Astronomia, exchange during galaxy merging events. (Assessing the Mass-Abundances Università degli Studi di Firenze, Italy The shape of the initial mass function redshift­ [Z] Evolution) is an ESO large 4 Jeremiah Horrocks Institute for of star formation also plays an important programme that was awarded 180 hours ­Astrophysics and Supercomputing, role, since different stellar masses inject of observations. It consists of deep ­University of Central Lancashire, into the interstellar medium different ­SINFONI, seeing-limited, integral field ­Preston, United Kingdom amounts of chemical elements. Clearly, spectroscopy of about 30 star-forming 5 Dipartimento di Astronomia, Università the metal content of galaxies is an im- galaxies (Lyman-break selected), most of di Bologna, Italy portant tracer of their star formation his­ which are at 3.0 < z < 3.7 and a few them 6 INAF–Osservatorio Astronomico di tory and of the main physical processes at 4.2 < z < 5.2. In the following we will Trieste,­ Italy involved in galaxy evolution. Indeed, the focus on the sample at z ~ 3.3. For these 7 Dipartimento di Fisica, Università di metallicity of local and distant galaxies galaxies the nebular lines [O ii] 3727 Å Trieste,­ Italy is one of the most important tools to con­ and [Ne iii] 3869 Å are redshifted into the 8 Graduate School of Science and strain galaxy evolutionary models. H-band, while Hβ and [O iii] 5007 Å are ­Engineering, Ehime University, Japan redshifted into the K-band. The flux ratio 9 Department of Physics and Astronomy, Clear observational evidence of a of these lines allows us to measure the University of California Los Angeles, ­connection between the content of met­ gas metallicity, as discussed in Maiolino USA als and star formation history is given et al. (2008). 10 INAF–Osservatorio Astronomico di by the tight three-dimensional correlation Bologna, Italy between metallicity, stellar mass and LSD (Lyman-break galaxies Stellar popu­ star formation rate (dubbed Fundamental lation and Dynamics) is a companion Metallicity Relation, FMR, Mannucci et ­programme that was awarded 70 hours The metal content in galaxies provides al., 2010), as illustrated in Figure 1. More of observations with SINFONI, assisted important information on the physical specifically, the gas metallicity is ob­­ by the adaptive optics module, so as processes responsible for galaxy for- served to increase as a function of the to achieve a much higher angular resolu­ mation, but little was known for galax- stellar mass (at a given star formation rate tion relative to the seeing-limited obser­ ies at z > 3, when the Universe was less [SFR]) and to decrease with the SFR (at vations. The sample consists of eight than 15 % of its current age. We report a given stellar mass). Galaxies show a Lyman-break galaxies at z ~ 3.3 selected on our metallicity survey of galaxies at very small metallicity scatter of 0.05 dex to have a nearby bright star, which is z > 3 using SINFONI at the VLT. We find around this surface and this suggests re­quired to guide the adaptive optics sys­ that at z > 3, low-mass galaxies obey that the bulk of galaxy formation occurs tem. the same fundamental relation between through a smooth, long-standing equi­ metallicity, mass and star formation rate librium between star formation, gas inflow In both projects the line emission (espe­ as at 0 < z < 2.5; however, at z > 3 mas- and outflow. Distant galaxies, out to cially the strongest one, [O iii] 5007 Å) sive galaxies deviate from this relation, z < 2.5, appear to obey the same funda­ is generally spatially resolved by our data.

36 The Messenger 142 – December 2010 9 2.2 0.8 9

8.75 2.2 2.2 0.8 1.4 8.75 2.2 H)

2.2 O/

H) 2.2

2.2 O/ 1.4 8.5 0.8 8.5

2.2 + log(

2.2 0.8 12 + log(

2.2 12 8.25 8.25 3.3 3.3 11 8 10.5 log(Ma 8 ss) 10 9.5 –1 11 1 0 –1 9.5 10 10.5 0 log(SFR) log(Mass) 1

Figure 1. Two views of the fundamental metallicity onds). Also in both samples extensive Troncoso et al., in prep.; Gnerucci et al., relation (FMR) between mass, star formation rate multiband photometry (including Spitzer in prep.), while follow-up observations are and metallicity for local galaxies are shown. The small points, with different colours, indicate different data) allowed us to constrain the stellar delivering additional results. It is beyond star formation rates. Squares with error bars indi- masses tightly. The star formation rate the scope of this paper to give an exten­ cate the average location of distant star-forming gal­ is inferred by using both the Hβ luminos­ sive overview of the various results. Here axies at different redshifts (the latter indicated by ity and spectral energy distribution (SED) we only show some of the main highlights the number associated with each point). The point at z ~ 3.3 deviating from the FMR was obtained with broadband fitting, generally obtaining that have been obtained by these pro­ the first preliminary set of 17 galaxies from AMAZE consistent results. grammes so far. and LSD. From Mannucci et al. (2010). A detailed description of these two pro­ grammes, as well as preliminary results, Chemical upsizing at z > 3 The projected spatial resolution is typically is given in Maiolino et al. (2008) and in about 5 kpc (seeing ~ 0.6 arcseconds) for Mannucci et al. (2009). Additional results The AMAZE and LSD samples span more the AMAZE data and about 1.5 kpc for have been, or are being published in five than two orders of magnitude in stellar the (nearly) diffraction-limited data in LSD additional papers (Gnerucci et al., 2010; mass (M ~ 109–1011 M ) and over an * A (point spread function [PSF] ~ 0.2 arcsec­ Cresci et al., 2010; Troncoso et al., 2010; order of magnitude in star formation rate

KNF,  ,   ~   KNF,  ,  Y  ~

 KNF,  ,   ~ '    ,~ .

Figure 2. Metallicity versus star formation rate in gal­   KNF  ,~  axies at 0 < z < 2.5 on the FMR (lines) and in galax­  ,~  ,  ies at z ~ 3.3 (squares) is shown for three different ~ Y]  stellar mass ranges. Within each panel different col­   ours give the metallicity values at exactly the same  ,   ,  ~ average stellar mass for galaxies both at 0 < z < 2.5 ~ and at z ~ 3.3. Note that while the metallicity of low- mass galaxies at z ~ 3.3 is consistent with local gal­ axies, massive galaxies at z ~ 3.3 are significantly    more metal-poor relative to their local counterparts.

KNF2%1,~XQ KNF2%1,~XQ KNF2%1,~XQ From Troncoso et al. (2010).

The Messenger 142 – December 2010 37 Astronomical Science Maiolino R. et al., AMAZE and LSD

Figure 3. [O iii] flux, velocity field and metallicity maps SSA22a−M38 for three massive galaxies at z ~ 3.3 characterised by regular rotation patterns. The metallicity has a 1.0 1.0 1.0 minimum close to the central peak of star formation 0.5 0.5 0.5 (as traced by the maximum of Hβ emission). From

Cresci et al. (2010). )

ǥ 0.0 0.0 0.0 ( R –0.5 –0.5 –0.5 (SFR ~ 30–300 MA/yr). Therefore, the metallicity inferred from the SINFONI –1.0 –1.0 –1.0 spectra allows us to obtain information on the mass–SFR–metallicity relation –0.7 0.0 0.71.4 –0.7 0.0 0.71.4 –0.7 0.0 0.71.4 (FMR) at z ~ 3.3. Figure 1 shows the loca­ [O III] Flux [O III] Velocity Metallicity tion of galaxies at z ~ 3.3 obtained from the average of the first set of 17 gal­ 0.0 0.20.4 0.6 0.81.0 –200 –100 0 100200 7.5 8.0 8.5 9.0 axies observed in AMAZE and LSD, Relative Intensity V (km/s) 12 + log (O/H) showing that galaxies at z ~ 3.3 clearly deviate from the FMR. SSA22a−M38

In Figure 2 (from Troncoso et al., 2010a) 1.0 1.0 1.0 we exploit the full AMAZE and LSD 0.5 0.5 0.5 ) ǥ final samples. The solid and dashed lines ( 0.0 0.0 0.0 show a cut of the FMR (i.e. the metal­ R licity–SFR relation) at three different val­ –0.5 –0.5 –0.5 ues of stellar mass. The AMAZE+LSD –1.0 –1.0 –1.0 data at z ~ 3.3 are shown with solid –0.7 0.0 0.71.4 –0.7 0.0 0.71.4 –0.7 0.0 0.71.4 squares that, for sake of clarity, give the [O III] Flux [O III] Velocity Metallicity average of the data in bins of mass and 9.2 SFR. Low-mass galaxies (M ~ 10 M , 0.0 0.20.4 0.6 0.81.0 –200 –100 0 100200 7.5 8.0 8.5 9.0 * A leftmost panel) have a metallicity in Relative Intensity V (km/s) 12 + log (O/H) line with the expectation of the relation observed at 0 < z < 2.5, indicating SSA22a−M38 that these low-mass galaxies at z ~ 3.3 are very much like local galaxies and 0.5 0.5 0.5 suggesting that they are regulated by the same evolutionary processes. However, 0.0 0.0 0.0 ) ǥ massive galaxies at z ~ 3.3, especially ( at M ~ 1010.7 M (rightmost panel), are R –0.5 –0.5 –0.5 * A significantly more metal-poor than galax­ –1.0 –1.0 –1.0 ies at 0 < z < 2.5 with the same SFR.

Taken at face value, this result seems to –1.5 –1.5 –1.5 imply that massive galaxies at z ~ 3.3 are in an earlier evolutionary stage rela­ –0.7 0.0 0.7 –0.7 0.0 0.7 –0.7 0.0 0.7 [O III] Flux [O III] Velocity Metallicity tive to their low-mass counterparts, in the sense that they have still to reach the metallicity–mass–SFR relation character­ 0.0 0.20.4 0.6 0.81.0 –300–200–100 0 100200 3007.5 8.0 8.5 9.0 ising galaxies at low-z. This is in con- Relative Intensity V (km/s) 12 + log (O/H) trast with the expectations of downsizing scenarios, where massive galaxies the central parts of massive galaxies, z ~ 3.3 in our sample, by measuring the should evolve faster and at earlier epochs hence diluting the metallicity of star- velocity shift of the brightest line, [O iii] relative to low-mass systems. forming regions. The latter scenario can 5007 Å. About 30 % of the galaxies in our be investigated by studying the dynam­ sample show ordered rotational motions There are a few possible scenarios ical properties of these systems, as in­ (Gnerucci et al., 2010a). A few examples that could explain the deviation of mas­ ferred by our SINFONI data. of such rotating systems are shown in sive galaxies at z ~ 3.3 from the FMR Figure 3, where the central panels show observed at 0 < z < 2.5. An excess of the rotation curve, along with the [O iii] pristine cold gas inflow in massive galax­ Massive rotating discs at z > 3 flux map (leftmost panels). ies, at such early epochs, may signifi­ cantly dilute the gas metallicity. Alterna­ The two-dimensional spectroscopic infor­ In relation to the chemical upsizing result tively, galaxy mergers may drive gas mation delivered by SINFONI allows us discussed above, we find that there is no from the outer, low metallicity regions into to trace the kinematics of the galaxies at correlation between the dynamical prop­

38 The Messenger 142 – December 2010 between dynamical and stellar masses 5L@W ѫHMS , 9$NAIDBSR (Mannucci et al., 2009; Gnerucci et al.,

5L@W ѫHMS +2#NAIDBSR 2010; Troncoso et al. in preparation).  Such high gas fractions are likely associ­ 5  L@W ѫHMS ated with the prominent cold inflows of 5  L@W ѫHMS gas predicted to occur at such early  epochs by some theoretical models (e.g., ] Y Dekel et al., 2009). 22  l# R "#% l" HD @W @K

2F 5L@WѫHMS Metallicity gradients and cold flows at  2(- "l" z > 3 "#%2l 22  l" R In some galaxies we not only resolve the

JL "#%2l 

S [O iii] 5007 Å emission (used to trace HM 5L@WѫHMS ѫ "#%2l the kinematics), but also the fainter lines of [O ii] 3727 Å and H , therefore en- R β @WHD F@K abling us to map the metallicity. The  +NB@K rightmost panels in Figure 3 (from Cresci 0l" 5  L@W ѫHMS et al., 2010) show the metallicity map for three massive galaxies characterised by regular rotation velocity fields. Surpris­ ingly, in contrast to local galaxies, the minimum metallicity is located close to the central regions. However, the most interesting result is that the minimum  "#%2l"#%2l 22  l, metallicity is associated with the peak of Hβ flux, which traces the most active     star-forming regions. This result supports the scenario where such massive sys­ 5L@WJLR tems at z ~ 3.3 drive major inflows of Figure 4. Velocity dispersion (σ) versus rotational ies, at z ~ 3.3 rotating discs are much pristine gas towards their central regions. velocity (V ) is shown for disc galaxies at z ~ 3.3 max more turbulent. Indeed, as illustrated in Such pristine gas both boosts star for­ with a regular rotation pattern. The solid line shows Figure 4, the velocity dispersion (σ) is mation and locally dilutes the pre-existing the relation Vmax/σ = 10 typical of local galaxies. Note that discs at z ~ 3.3 are much more turbulent than generally comparable with the rotational medium, therefore producing the local galaxies, many of them having V / ~ 1. From max σ velocity (Vmax). More specifically, the observed spatial anticorrelation between Gnerucci et al. (2010). ­average ratio between rotational velocity star formation and gas metallicity.

and velocity dispersion is z = 3.3 = 2.2, to be compared with the value The same (strong) cold flow scenario can erties of massive galaxies at z ~ 3.3 and z = 0 = 10 typical of local discs. explain, more generally for massive their deviation from the fundamental Galaxies at z ~ 3.3 appear to be even ­galaxies at z ~ 3.3, their reduced metal­ metallicity–mass–SFR relation observed more turbulent that those investigated at licity relative to the fundamental metallic­ at 0 < z < 2.5. More specifically, among z ~ 2 by the parallel SINFONI programme ity–mass–SFR relation observed at the z ~ 3.3 massive galaxies, which SINS (Förster Schreiber et al., 2009), which 0 < z < 2.5, as well as the highly turbulent are metal-poor relative to the metallicity– are characterised by z = 2 = 4.5. nature of these systems. mass–SFR relation at 0 < z < 2.5, we find systems with both irregular kinemat­ The highly turbulent nature of z ~ 3.3 ics (likely tracing merging systems) and discs is likely due to very high gas References a regular rotation curve, in equal num­ ­fractions, which make the discs dynami­ Cresci, G. et al. 2010, Nature, 467, 811 bers. Therefore, merging cannot be the cally unstable. In samples at 1 < z < 2.5 Daddi, E. et al. 2010, ApJ, 713, 686 only process responsible for lowering high gas fractions have been con- Dekel, A. et al. 2010, Nature, 457, 451 the gas metallicity in these early galaxies. firmed directly through CO observations Förster Schreiber, N. M. et al. 2009, ApJ, 706, 1364 (Daddi et al., 2010; Tacconi et al., 2010). Gnerucci, A. et al. 2010, A&A, submitted, arXiv 1007.4180 From the rotation curve we can also infer At z ~ 3.3 we have obtained indirect Maiolino, R. et al. 2008, A&A, 488, 463 dynamical masses, which are in the ­evidence for high gas fractions (even Mannucci, F. et al. 2010, MNRAS, 408, 2115 9 11 Mannucci, F. et al. 2009, MNRAS, 398, 1915 range 2 × 10 – 2 × 10 MA. Clearly, some approaching fgas = Mgas/Mtot ~ 0.9) based massive rotating discs are already in on the high surface density of star for­ Tacconi, L. J. et al. 2010, Nature, 463, 781 Troncoso, P. et al. 2010, A&A, submitted place at this early epoch in the Universe. mation (hence exploiting the Schmidt– However, in contrast to local disc galax­ Kennicutt relation) and on the comparison

The Messenger 142 – December 2010 39 Astronomical News and Roger Angel. from left to right, Nelson, Jerry Raymond Wilson to, Astrophysics in Prize 2010 Kavli the presents right) (far Norway of Harald King Majesty His article by Filippi (p. 2). (p. Filippi by article the 2010. See 4November on Chile Santiago in Offices ESO at held project network speed high Access to Latin-American Southern Observatories) Virtual (Enabling EVALSO the of inauguration The

Credit: Scanpix Astronomical News

Raymond Wilson Honoured with Two Prestigious Prizes

Jeremy Walsh1

1 ESO

Ray Wilson, who retired from ESO in 1993, was awarded two prestigious prizes in September 2010 for his out­ standing work on telescope optics: the Kavli Prize in the field of astrophysics and the Tycho Brahe Prize 2010 of the European Astronomical Society.

Announced in June 2010, and awarded in Stockholm in September, the million- dollar Kavli Prizes were awarded to eight scientists “whose discoveries have ­dramatically expanded human under­ standing in the fields of astrophysics, nano­science and neuroscience” (see the Kavli press release1). The Kavli prize is awarded by the Norwegian Academy of Science and Letters, the Kavli Foundation and the Norwegian Ministry of Educa- tion and Research. The Kavli Foundation VLT and the 42-metre ELT: the develop­ Figure 1. Ray Wilson receiving the Tycho Brahe prize is funded by Fred Kavli, the Norwegian ment of active optics as the basis of from the president of the EAS at the JENAM meeting in Lisbon, Portugal. entrepreneur and philanthropist who later all modern telescope optics”. Figure 1 founded the Kavlico Corporation in shows Ray receiving the Tycho Brahe the US — today one of the world’s largest prize from the retiring president of the During his last years at ESO he began suppliers of sensors for aeronautical, EAS, Joachim Krautter. work on his magnum opus, the two-vol­ automotive and industrial applications. ume work Reflecting Telescope Optics, There were three recipients in astro­ Ray Wilson, who was born in England published by Springer 5. Volume I: Basic physics, all acclaimed for their work on and educated at Birmingham University Design Theory and Its Historical Develop- the development of giant optical tele­ and Imperial College London, arrived at ment, first appeared in 1996, and Volume scopes2 — Roger Angel of the University ESO in 1972 from Zeiss at Oberkochen, II: Manufacture, Testing, Alignment, Mod- of Arizona, Tucson, attached to Steward Germany where he had been head of the ern Techniques followed in 1999. Both Observatory; Jerry Nelson of the Uni­ Optical Design Department for Astro­ are currently in their second edition and versity of California, Santa Cruz and long nomical and Analytical Instruments. At Ray is working on updates for the third associated with the Keck Observatory; ESO, first in Geneva and then in­ Garching, editions. Ray has also been honoured by and Ray Wilson. The photograph on he was the first head of the ESO Optics a number of other prizes, including the the facing page (upper) shows the three and Telescopes Group. The revolutionary Karl Schwarzschild Medal of the German prize winners at the prize award cere­ active optics of the 3.58-metre NTT, in- Astronomical Society, an appointment as mony in Oslo. spired by two years of work at the La Silla Chevalier of the French Légion d’Honneur Observatory, was the crowning achieve­ and the Prix Lallemand of the French The Tycho Brahe Prize of the European ment of Ray’s work at ESO. This success­ Academy of Sciences. Astronomical Society (EAS) is awarded ful concept, where both the alignment of annually in recognition of the develop­ the optical elements as well as the shape ment or exploitation of European instru­ of the flexible primary mirror are con­ Links ments, or major discoveries based large-­ trolled in a closed loop based on the 1 The Kavli Prize: http://www.kavlifoundation.org/ ­ly on such instruments. The prize is measurements of a wavefront sensor, kavli-prize sponsored by the Klaus–Tschira founda­ was then also used for the VLT 8.2-metre 2 Announcement of Kavli Prize for astrophysics: tion, based in Heidelberg, Germany. An- telescopes. In addition, Ray contributed http://www.kavlifoundation.org/2010-astrophysics- nounced in April, Ray received the prize to telescope designs with more than citation 3 ESO announcement of Kavli Prize: at the JENAM meeting in Lisbon, Portu­ two powered mirrors, which are now http://www.eso.org/public/news/eso1022/ gal (see the article by Sandu and Chris­ being explored for the next generation of 4 Announcement of Tycho Brahe Prize: tensen, p. 42) and, at the plenary session extremely large telescopes, such as http://www.eso.org/public/announcements/ on 10 September 2010, he delivered a the European Extremely Large Telescope ann1018/ 5 Wilson, R. N. 2004, Reflecting Telescope Optics, lecture entitled “From the ESO NTT to the (E-ELT) project. Volumes I and II, 2nd edition, (Heidelberg: Springer)

The Messenger 142 – December 2010 41 Astronomical News

Availability of Reduction Software for HARPS Data at ESO Headquarters in Garching

Gaspare Lo Curto1 initial guess for the radial velocity). In (e.g., introduction to the data reduction Thierry Beniflah1 these cases the spectral extraction is not system, etc.). Regrettably ESO will Andrew Burrows1 affected, but the precision and the accu­ not be able to cover any expenses (travel, Eric Emsellem1 racy of the radial velocity measurement ac­commodation, etc.) for these “data Kevin Maguire1 are generally not optimal. Although these reduction missions”. The visitor will use Luca Pasquini1 cases are not frequent, they do happen her/his own laptop to run the data reduc­ John Pritchard1 from time to time and require re-compu­ tion software remotely via the standard Martino Romaniello1 tation of the radial velocities. We aim to GUIs. The raw data will then be trans­ address such needs by allowing indi­ ferred to the reduction machine, either vidual users to visit ESO Headquarters in from a laptop (i.e. via ftp) or from the 1 ESO Garching and give them access to the archive ftp site after an archive request same data reduction software that is has been processed. Saving the reduced available at the La Silla Observatory site, data is the responsibility of the user. From the start of the year 2011 the HARPS both in its on- and off-line flavours. ­Visitors are expected to spend at most data reduction software will be also avail­ five days on each data reduction mission, able at ESO Headquarters in Garching. Users wishing to take advantage of this and the service will be available during This new initiative will enable users to ap- service are encouraged to check the normal office hours from Monday to ply for access to the system locally in details at the web page1. The user should ­Friday at ESO Headquarters in Garching. Garching. then contact ESO giving a brief scien- tific and technical rationale as to why Users wishing to employ this HARPS ­re- The experience of the past few years of reprocessing is required, together with duction service in Garching should send operations of HARPS has shown that the amount of data that needs to be an email to [email protected]. sometimes the data reduced online at the reduced and the intended dates of travel La Silla Observatory might require further to Garching. ESO, after checking availa­ reprocessing and analysis. This is usually bility, will make available desk space and Links because the wrong set of initial parame­ will grant access from the user’s laptop 1 Details of the service at: www.eso.org/sci/facilities/ ters have been specified in the observing to the data reduction computer. Limited lasilla/instruments/harps/tools/reprocess.html template (such as stellar spectral type or on-site user support will also be provided

ESO Participation at the Joint European and National Astronomy Meeting in Lisbon, Portugal

Oana Sandu1 ESO was extensively involved in the Tamai, Head of ESO’s Technology Divi­ Lars Lindberg Christensen1 meeting, highlighting its role as a driving sion, showcased the technology of the force in ground-based astronomy at the VLT/VLTI, while Roberto Gilmozzi, Head European level, as well as globally. Sev­ of ESO’s Telescope Division, presented 1 ESO eral key ESO people participated in the principal technological features of the the meeting and there were also an ESO European Extremely Large Telescope ­plenary and special session, an ESO (E-ELT); a project that incorporates many The Joint European and National Astron­ exhibition with free educational and infor­ innovative developments. A presentation omy Meeting (JENAM) that took place mational material and a book launch. on control software and data reduction in Lisbon, Portugal, during the week of and analysis was delivered by Michèle 6–10 September 2010, was the 18th During the first day of the meeting, ESO Péron, Director of Engineering and Soft­ Annual Meeting of the European Astro­ participated in the special session on ware Development. nomical Society (EAS) and the 20th Astronomy Challenges for Engineers and Annual Portuguese Meeting of Astronomy Computer Scientists with talks by Bruno On Tuesday, 7 September, there was a and Astrophysics. JENAM brings Euro­ Leibundgut, Director for Science, on dedicated ESO plenary session. Bruno pean astronomers together to discuss ­science projects at ESO and Andreas Leibundgut gave a comprehensive talk frontline topics in astronomy, space sci­ Kaufer, Director of La Silla Paranal Obser­ about recent developments at the La Silla ence and instrumentation technology. vatory, on ESO’s infrastructures. Roberto Paranal Observatory, including plans for

42 The Messenger 142 – December 2010 Figure 1. The ESO stand at JENAM in Lisbon, Portu­ gal shown during the launch of the the book Post- cards from the Edge of the Universe. Catherine Cesarsky, former ESO Director ­General and past President of the IAU, is shown discussing the book. the E-ELT. He was followed by the lead- the instrument that will be the world’s Cesarsky. A book signing session took er of the ESO Survey Team, Magda biggest eye on the sky. place and free postcards signed “Greet­ ­Arnaboldi, who presented the status of ings from the Edge of the Universe” the Public Surveys that have recently In order to encourage interaction be- were distributed to people, who were begun with VISTA. tween participants and foster debates on encouraged to post them to their family topics related to astronomy in a more and friends. More infor­mation on obtain­ A special session on ALMA Early Science relaxed environment, ESO ran the ESO ing PDF copies of the book, or infor­ took place on the same day. The session Hour at its stand on three days of the mation on how to order a hard copy or included presentations on the current meeting. The event proved to be suc­ send electronic postcards, are available 2. status of the construction project, the cessful, gathering numerous participants ALMA development plan and its opportu­ from the meeting who engaged in con­ Among other highlights of the meeting nities, and the European ALMA Regional versation with people from the organisa­ included the inauguration of the EAS Centre plans for user support in prepa­ tion. Lodewijk Woltjer Lecture award talks with ration for Early Science. The session also Woltjer himself giving the first talk. Under included a demonstration of the ALMA One of the highlights at the ESO stand his leadership as Director General, ESO software that will be used to apply for ob- was the launch of the book Postcards signed the VLT contract and established serving time on ALMA, to prepare obser­ from the Edge of the Universe. This book itself as one of the world’s leading astro­ vations, interact with the ALMA archive is based on the science carried out by nomical institutes. Lodewijk Woltjer also and reduce the data. a hand-picked selection of the best blog­ made significant contributions to theore­ gers from the Cosmic Diary1, one of the tical astrophysics from his fundamental ESO also participated in the session ded­ 12 Cornerstone projects of the Interna­ work on the Crab Nebula and his studies icated to Education and Outreach after tional Year of Astronomy 2009. Twenty- on the energy source of radio galaxies IYA2009 in Europe. Pedro Russo, Global four astronomers from all corners of and quasars. Another highlight of the Coordinator for the International Year of the globe explain their science in articles JENAM meeting was the award of the Astronomy 2009, talked about his experi­ edited by Lee Pullen, Mariana Barrosa EAS Tycho Brahe Prize for 2010 to tele­ ence in spearheading this project. and Lars Lindberg Christensen. scope designer Ray Wilson (see the arti­ cle on p. 41). During the five days of the conference, Announced several days in advance ESO exhibited its astronomical dis­ and kept as a surprise, the book launch coveries, and the telescopes that made took place on Tuesday, when several Links these discoveries possible. The E-ELT of the authors and editors talked about 1 Cosmic Diary: http://www.cosmicdiary.org project attracted the interest of many the book (see Figure 1). A special talk 2 Postcards from the Edge of the Universe: passers-by who admired the telescope was given by a guest of honour — the http://www.postcardsfromuniverse.org model and went on to learn more about former ESO Director General Catherine

The Messenger 142 – December 2010 43 Astronomical News

Visiting ESO’s Office in Santiago

Michael West1

1 ESO

Astronomers travelling to Chile to observe with ESO telescopes are encouraged to include a visit to ESO’s office in Santiago. With the recent completion of the new ALMA building (see Figure 1), there is now a vibrant scientific community of more than 100 as­­tronomers, fellows and stu­ dents at the ESO premises in Santiago. ESO’s Office for Science in Chile is happy to provide accommodation and per diem for visiting astronomers who wish to give a science talk and interact with ESO and ALMA staff. If interested, please contact Michael West ([email protected]). We look forward to welcoming you to Santiago!

Figure 1. The hand-over ceremony for the new ALMA offices at the ESO premises in Santiago. Tim de Zeeuw (ESO Director General, left) and Thijs de Graauw (ALMA Director, right) exchange the key.

Café & Kosmos Events in Munich

Henri Boffin1 Hannelore Hämmerle 2 Barbara Wankerl3 Silke Zollinger4 Credit: A. Griesch/MPP

1 ESO 2 Max-Planck-Institut für extraterrestrische Physik, Garching, Germany 3 Excellence Cluster Origin and Structure of the Universe, Garching, Germany 4 Max-Planck-Institut für Physik, Munich, Germany

On a Monday evening such as the 5th of July this year, the trendy bar Café Jasmin in the Schwabing district of Munich is full inside, despite the hot air and blazing Figure 1. An interested crowd in lively discussion Sun that would normally invite people with the physicist Stefan Stonjek from MPP at the to enjoy a beer outside in a tree-shaded first Café & Kosmos event.

44 The Messenger 142 – December 2010 Astronomical News

garden. Although we are in the midst The idea is to bring science directly to people the chance to speak directly with of the international football season, these the general public in the Munich area in scientists about current fascinating sci­ people did not gather to watch another the relaxed atmosphere of a bar. entific themes. game of the World Cup; oddly enough, these are members of the public inter­ For many, research is “far away” and The proposed themes for Café & ested in science, and in our Universe in happens behind the closed doors of Kosmos come from astrophysics, cos­ particular. They are here to listen and laboratories. Admittedly, there are many mology and particle physics, and are discuss with ESO astronomer Markus ­science magazines in print or on televi­ on topics of great interest for non-scien­ Kissler-Patig, the fundamental question: sion and some newspapers have a page tists, such as “What are black holes?”, Are we alone in the Universe? about science, but the contact between “What do we know about dark matter?”, science and the public is often very in­ “Why did CERN build the LHC?”, “What Subsequent science cafés have con­ direct. Many scientific institutions organ­ do we know about planets outside our firmed the success of the first event ise open days, which are often very Solar System?”, and so on. The discus­ that took place on 31 May 2010, when ­successful, but these take place only sions are held in German. These and Stefan Stonjek, a physicist from the once a year at most. Public conferences other topics are discussed in the relaxed Max-Planck-Institut für Physik (MPP), on scientific themes are also often popu­ atmosphere of a pub — every first Mon­ ­discussed The Big Bang in the Tunnel, lar, but they generally follow the same day of the month. The duration of the covering the latest details about the academic scheme: the scientist speaks ­discussions is initially about one and a CERN Large Hadron Collider (see Fig- and the public listens. Direct exchanges half hours, although our first two experi­ ure 1). The complex topic of string theory between scientists and the public seldom ences have shown us that many people was aired by Ilka Baumgartl and Marco take place. tend to stay for much longer. Baumgartl from the Excellence Cluster in September and the subject of the black The Café & Kosmos initiative aims to For more information on the Café & hole at the heart of the Milky Way was bring researchers and non-scientists Kosmos series, including the discussed by Stefan Gillessen from the together, and to do so in places where list of future speakers, please go to Max-Planck Institute for Extraterrestrial people typically meet, share their http://www.cafe-kosmos.org. Physics in November. The series is enti­ thoughts, discuss business and debate tled Café & Kosmos and is a joint initiative about big and small things. Thus a pub It is a pleasure to thank Aleks Vulic, between ESO, the Excellence Cluster in the centre of the city of Munich was owner of Café Jasmin, for his permission ­Origin and Structure of the Universe, and chosen for the meetings; a place where to use his premises for these interactive the Max-Planck Institutes for Physics, communication traditionally takes place. sessions, as well as the speakers. Astrophysics and Extraterrestrial Physics. With Café & Kosmos, we want to give

New Staff at ESO

Jean-Philippe Berger my passion for history and astronomy barely capable of flying to the planets in occurred in a remote Spanish village the Solar System. Then at the age of 13 I I have been at ESO as a VLTI staff as­ when I was approximately ten years old. received a book from my uncle: Le Ciel tronomer, on leave from the Laboratoire My uncle had a TV (our family did not) by Jean-Claude Pecker. This was the d’Astrophysique de Grenoble (France), and I watched an episode of the well- true revelation that one could “travel” in for about 10 months. I came to Santiago known science fiction series Cosmos space thanks to telescopes and an imag­ with my wife Stephanie and our three 1999. It came as a revelation that inative brain. children Clara, Lucie and Axel. man could travel into space and reach remote parts of the Universe. I imme­ Even though I continued to follow his- I remember my very early fascination with diately became interested in rockets and tory lectures throughout my academic ruins that later turned into a deep inter- space shuttles. Finding and studying career, my educational path slowly est in human civilisations and especially new civilisations seemed to be a realistic but surely shifted towards science and their dawn and dusk. This has never project and it took me a few years technology and the final call to become left me. I believe the connection between to accept that current technology was an astronomer came during my short

The Messenger 142 – December 2010 45 Astronomical News

stay at ENS Lyon for my masters in astro­ Jean-Philippe Berger physics. After a two-year stay in Chad, Africa, as a physics teacher, I joined the Laboratoire d’Astrophysique de Grenoble (LAOG) in 1995 for a PhD on the polar­ ising properties of dusty environments in young pre-main sequence binaries. Towards the middle of my PhD I realised that I wanted to become involved in an instrumentation project. One day, on the stairway, I met a colleague who was also an optical engineer, Pierre Kern, who was very excited by his new idea to use the miniature optical circuits developed by the telecommunications industry to produce interference between the light of two telescopes. I was immediately struck by the potential of this technology for aperture synthesis in the optical and asked him if there was a way for me to contribute to the project. Unfortunately he was applying for a grant to recruit a PhD student and had a good candidate; but fortunately (for me) the grant never I continued developing integrated optics Université Joseph Fourier and in public arrived and I spent the remaining part of technology after my thesis, thanks to ­outreach and realised that it was an es­ my PhD developing this new and prom­ the support of the French Space Agency sential part of our mission as “astron­ ising technology. Since then the term (CNES). I dived into the photonics world omes de la République”. Finally, in 2008, “Astrophotonics” has been coined to for two years at LEMO and in 2000 I at a conference in Marseille, some of describe the marriage between photonics obtained a NASA/JPL Michelson Fellow­ my colleagues and myself were impressed and astrophysics. ship to work with Wesley Traub and his by the imaging results coming out from team at the Harvard–Smithsonian Center the MIRC instrument at CHARA and So, eighteen months from the end of my for Astrophysics. My project was to install became convinced that the VLTI and PhD grant I started everything from an instrument, IONIC3, capable of com­ integrated-optics technology were ready scratch, and I found myself alone in an bining three beams of the IOTA interfer­ for four-telescope operation. This dis­ empty lab with the task of exploring ometer in Arizona and to use it to gener­ cussion gave birth to the PIONIER VLTI how to apply photonic technology to ate the first images of the circumstellar visitor instrument project that I initiated astronomical interferometry. Fortunately I environment in a pre-main sequence star. and led in collaboration with some in married Stephanie almost at the same I remember those American years with December 2008. Since then PIONIER has time as I changed my research path and a lot of affection, and the seemingly been built and it received its first stellar her presence by my side for all these countless days and nights spent at IOTA photons in October 2010. years has been essential in holding me to (Mount Hopkins, Arizona) and CHARA this new path. My thesis culminated in (Mount Wilson) still live with me. We Coming to ESO and VLTI was thus a resolving the accretion disc of a young expended a huge amount of energy to try ­natural move. First I could follow PIONIER star, FU Orionis, for the first time with to produce the first maps of young stellar and secondly I was keen to join the astronomical unit resolution, working with objects in the near-infrared. This was impressive Paranal astronomical machine my unofficial but remarkable advisor ­partially successful, but clearly limited by and the excellent VLTI team. and friend Fabien Malbet, who taught me the lack of long baselines at IOTA. how to use the Palomar Testbed Inter­ ferometer, which was then the only instru­ In 2002 I took up a permanent position ment with sufficient sensitivity to study at the Observatoire de Grenoble, where protoplanetary discs. At the end of my I continued to develop new projects PhD I realised that my research interests related to astronomical aperture synthe­ combined challenging astrophysics with sis imaging. I also started to get more challenging instrumentation. and more involved in teaching at the

46 The Messenger 142 – December 2010 Astronomical News

Fellows at ESO

Margaret Moerchen Margaret Moerchen

I grew up in central Texas, where city lights frequently obscured the nightly show of constellations. (While we proudly refer to Texas as The Lone Star State, it isn’t for this reason!) When it was time for a meteor shower, I could make a short drive into the hill country where a much richer sky was revealed — and perhaps it’s because the stars weren’t always in plain sight that these glimpses were so inspiring.

These night-time excursions continued while I did my undergraduate course- work at the University of Texas at Austin. I then found an even darker sky in the Davis Mountains, home of McDonald later, I’m back at the VISIR console as I often ask myself, how many decisions Observatory, where I had my first experi­ instrument fellow and am part of the team I made to get here. Until I was more than ence using a “big” (0.75-metre) telescope. working on its exciting upgrade project. half way through my university physics After graduating, I took my first step My functional duties include support of degree I was not thinking of being a pro­ into mid-infrared astronomy by working observations not only with VISIR, but with fessional astronomer. I had not even on the design of EXES (Echelon Cross all instruments at Unit Telescope 3 (UT3) looked through a telescope until about Echelle Spectrograph), a spectrograph and UT4, and at UT4 I’ve been able to that time. One of the first convincing that will soon fly on SOFIA, the Strato­ learn more about adaptive optics and the moments was when I joined a group of spheric Observatory for Infrared Astron­ near-infrared regime. Performing a wide fellow students on a visit to the observa­ omy. variety of science programmes at the two tory in the sleepy town of Višnjan in Croa­ telescopes has provided a unique per­ tia, where amateur astronomers were Wanting to continue in the hot field of the spective on the power of 8-metre tele­ becoming professionals in their achieve­ thermal infrared, I went to the University scopes, and welcoming visiting astrono­ ments of spotting Near Earth Objects. of Florida to participate in the building of mers and hearing enthusiastic detailed There I had a first glimpse of the life of an mid-infrared instruments such as T-ReCS descriptions of their projects is one of the astronomer: it was not just star-gazing. (Thermal Region Camera and Spectro­ most rewarding aspects of the support There was a lot of careful and patient tak­ graph) for Gemini South. I spent five work. In fact, even driving astronomers ing of pictures, comparing them, and monsoon seasons in Gainesville, where up or down the mountain in the middle of working with different image-processing my thesis research employed T-ReCS the night offers one of the benefits we software, and so on, throughout the and other mid-infrared cameras to study sometimes forget — to look at the night night, until the rain came. It was fun! the architecture of young planetary sys­ sky! I’ve seen the most spectacular tems by determining the location of warm crumbling-fireball meteors of my life while The starting point in my career was the dust within them. At the same time, I had on the summit road, and they’re an im­ acceptance to do a PhD at the Leiden the opportunity to become involved in pressive reminder of why I travelled thou­ Observatory. Very early on, I went to the development of CanariCam (presently sands of miles to stay up at night in the observe with the SAURON integral-field in the commissioning phase), the mid-IR middle of the desert. unit mounted on William Herschel Tele­ facility instrument for the Gran Telesco- scope on La Palma. The fifth observing pio Canarias, the 10.4-metre segmented- run of the SAURON Project was 14 nights mirror telescope on La Palma. Davor Krajnović long and a number of team members came to share the time, but I stayed for After attending the Observing Planetary On a warm summer night, in the com­ the full run. Two weeks on the mountain, Systems workshop at ESO Chile in 2007, pany of a dentist, a physician and a man­ not having to close the dome once, I thought this would be a fantastic place ager, in the cockpit of a sailing boat that catching as many photons as the detec­ to combine the pursuits of both scientific rocks slowly, while a breeze brings the tors allowed, from dusk until dawn, and, research and instrumentation develop­ smell of pine trees and the buzzing sound just before going to sleep, looking at the ment. I became even more sold on ESO of a few persistent cicadas, and the stars shadow of the mountain on the clouds after I had the chance to visit Paranal shine on us with the intensity of a crystal- or the Atlantic Ocean, this was something to assist in a post-intervention characteri­ clear night, the dentist concludes: “What extraordinary. I was hooked. sation of VISIR. Now, almost three years a good job you have: looking at the stars!”

The Messenger 142 – December 2010 47 Astronomical News

I spent four wonderful years in Leiden. observers and theoreticians. The first My thesis consisted of the analysis of results are coming out right now and it the nuclear properties and dynamical is amazing to see how the initial ideas modelling of nearby early-type galaxies, have turned into science. as well as developing a method to ana­ lyse the two-dimensional kinematic It is now about ten years since I left my maps coming from instruments such as hometown and I am entering my sec- SAURON. After the PhD, I went to Oxford ond year as an ESO fellow. Astronomy where I stayed for five years. For the really is not just star-watching, and ESO last two I was an Extraordinary Junior is a prime example of the complexity Research Fellow of Queen’s College (the needed for successful astronomical oper­ “Extra” actually means they were not ations. I feel rather privileged to be able paying my salary, but they did take care to participate in it. On a warm summer of me as a college does). There I started night, however, when a friend says it is a working in the near-infrared and with good job, this star watching, one has to laser-guided adaptive optics observa­ make a decision: to agree, or say: “Well, tions, which unfortunately also meant that actually I study black holes ...” trips to the telescopes became rarer. A large part of my science, however, con­ centrated on a new and exciting project, the ATLAS3D survey, which I am co- leading. This is a multi-wavelength survey of a complete sample of nearby early-type galaxies, and it includes a large team of Davor Krajnović

In Memoriam Christine Nieuwenkamp

Tim de Zeeuw1 Through her strong motivation and dedi­ cation, Chris was able to cope with a constantly high workload, especially 1 ESO related to calls for tender, contracts and contract administration, and inter­ acted with a large number of staff in Christine (Chris) Nieuwenkamp was born both Garching and Chile. In addition in Belgium and studied at the Higher Chris successfully trained and integrated Institute for Translators in Antwerp. She new staff in the Contracts and Procure­ started her professional career in the ment Department. For several years purchasing department of a Brazilian min­ Chris was an active and respected mem­ ing company stationed in Belgium. She ber of the ESO Staff Association in joined ESO in April 1990 as Administrative Garching. Her excellent performance as Assistant for Purchasing in the Contracts a Procurement Officer was well-recog­ and Procurement Service at ESO Head­ nised both inside and outside ESO and quarters in Garching. she was a pleasure to work with. She died on 18 October 2010. Chris will be remembered by her friends and col­ leagues at ESO and missed by her hus­ band and two children.

48 The Messenger 142 – December 2010 Astronomical News

Announcement of the ALMA Community Days: Towards Early Science

6–7 April 2011, Garching, Germany

ALMA, the Atacama Large Millimeter/ Science Operations in the second half of With the ALMA Community Days, the submillimeter Array, is expected to be the 2011. During the first phase of Early ESO ARC aims to prepare the European leading observatory at millimetre and ­Science, an array of 16 antennas will be astronomical community for ALMA submillimetre wavelengths over the com­ offered for interferometry with four fre­ Early Science operations. The first day ing decades. It is the result of a global quency bands and a limited range of will be dedicated to a series of scientific collaboration involving Europe, North baselines. Early Science observations are and technical presentations related America, East Asia and the host country currently estimated to be scheduled for to ALMA and Early Science capabilities, Chile. When completed, it will com- at most one third of the available time, while the second day will be taken up prise at least 66 high precision antennas the remainder being reserved for continu­ by interactive tutorials on the preparation equipped with receiver and digital elec­ ing commissioning and science verifica­ of ALMA observing proposals using the tronics systems to observe in the fre­ tion activities. ALMA Observing Tool (OT). This should quency range from 30 GHz to 1 THz and help novice and advanced ALMA users achieve angular resolutions as high as Scientific users will interact with the alike to create observing projects that 5 milliarcseconds. Dynamic scheduling ALMA facility through their local ALMA optimally exploit the unique capabilities of and innovative calibration strategies Regional Centre (ARC), which will provide ALMA during Early Science operations. will ensure the most efficient use of the user support on all aspects related to unique atmospheric qualities encoun­ observing with ALMA and assist observer Further information can be found at tered at the 5000-metre high site on the teams throughout the lifecycle of their www.eso.org/sci/meetings/2011/ Chajnantor plateau in the northern Andes. project. The European ALMA community alma_es_2011.html or by emailing is supported by a network of regional [email protected]. While Full Science Operations are esti­ ARC nodes that are coordinated by mated to begin in 2013, the increasing the central European ARC hosted at ESO capabilities of the growing array will Headquarters in Garching, Germany. become available to the astronomical community following the start of Early

Announcement of the ESO Workshop Fornax, Virgo, Coma et al.: Stellar Systems in Nearby High Density Environments

27 June–1 July 2011, Garching, Germany

This workshop will provide an overview of Topics to be covered include: Scientific Organising Committee: recent observational results on the stellar 1) the faint end of the galaxy luminosity Magda Arnaboldi (chair), Nobuo Arimoto, systems in nearby galaxy clusters, i.e. function in clusters: dwarf galaxies, Michele Cappellari, Eric Emsellem, Bill Fornax, Virgo, Coma et al., and a forum for ultra-compact dwarfs and globular clus­ Harris, Ken Freeman, Ortwin Gerhard, discussion and comparison of theoretical ters; John Kormendy, Harald Kuntschner, models for the evolution of galaxies and 2) the bright end of the galaxy luminosity Claudia Maraston, Lucio Mayer, Simona larger-scale structures with observational function: stellar populations and dynam­ Mei, Bianca Poggianti, Sadanori Okamura, properties of stellar systems in high den­ ics — observation and theory; Tom Richtler. sity environments at redshift zero. The 3) surveys of nearby clusters and their fol­ aim is also to identify those questions that low-up with the ESO Very Large Tele­ Local Organising Commitee: can be tackled by the European Extremely scope (VLT) and future facilities (E-ELT, Magda Arnaboldi, Lodo Coccato, Luca Large Telescope (E-ELT), as the excep­ Atacama Large Millimeter/Submillimeter Cortese, Michael Hilker, Marina Rejkuba, tional high angular resolution and collect­ Array, ALMA); and Christina Stoffer. ing power of an extremely large telescope 4) the 3D structure of nearby galaxy clus­ are essential ingredients for the study ters and the morphological transforma­ Web page for registration and further of resolved stellar populations at distances tion of galaxies. details will appear soon at http://www.eso. larger than 10 megaparsecs. The nearby org/sci/meetings.html. clusters Fornax, Virgo, Coma et al. will The number of participants is limited be the first obvious targets of this exciting to the capacity of the ESO auditorium new era in extragalactic astronomy. (about 90 participants).

The Messenger 142 – December 2010 49 Astronomical News

Personnel Movements

Arrivals (1 October–31 December 2010) Departures (1 October–31 December 2010)

Europe Europe Ascenso, Joana (P) Fellow Beniflah, Thierry (F) Head of Information Technology Beccari, Giacomo (I) Fellow Bois, Maxime (F) Student Boissier, Jérémie (F) Fellow Cortesi, Arianna (I) Student Bressert, Eli (USA) Student Frank, Matthias (D) Student Coccato, Lodovico (I) Fellow Guidetti, Daria (I) Student Gabasch, Armin (I) Software Engineer Hansen, Camilla Juul (DK) Student Heidecke, Thies (D) Student Hewitson, Jennifer (D) Secretary/Assistant Kains, Noé (B) Fellow Larsen, Jonas (DK) Software Engineer Lablanche, Pierre-Yves (F) Student Lombardi, Marco (I) Astronomer Longmore, Steven (GB) Fellow Naets, Thomas (B) Internal Auditor Lützgendorf, Nora (D) Student Neumayer, Nadine (D) Fellow Pettazzi, Lorenzo (I) Control Engineer Penuela, Tania Marcela (CO) Student Ruiz Velasco, Alma (MEX) Student Russo, Pedro Miguel (P) Astronomer Seemann, Ulf (D) Student Taylor, Luke (GB) Laser Physicist Unterguggenberger, Stefanie (A) Student Williams, Michael (GB) Student Chile Chile Alvarez, Fernando (RCH) Head of the Maintenance Department Asmus, Daniel (D) Student Gonzalez, Jaime (RCH) Optical Technician Kaminski, Adrian (D) Student Jones, Matias (RCH) Student Raffi, Gianni (I) Head of the Computing IPT Miccolis, Maurizio (I) System Engineer Thomas, Alexis (RCH) Network Specialist Rodrigues, Myriam (P) Fellow Varas, Oscar (RCH) Dome Mechanic Vanderbeke, Joachim (B) Student Vega, Florine (RCH) Procurement Clerk Credit: ESO/Y. Beletsky

A wide field image of the southern hemisphere con­ This OB association was studied by Adriaan Blaauw stellations Crux and Centaurus which make up part in his PhD thesis (1946) and published in Bulletin of the Scorpius Centaurus Association (Sco OB2), of the Astronomical Institutes of the Netherlands the nearest region of recent massive star formation. (Blaauw 1952, BAN, 11, 414).

50 The Messenger 142 – December 2010 Astronomical News

In Memoriam Adriaan Blaauw

Adriaan Blaauw, the European Southern research organisation that could com- Observatory’s second Director General, pete on the international level. Blaauw and a key figure in ESO’s early history, had returned to Leiden in 1948, moved to died on 1 December 2010, at the age Yerkes Observatory in 1953, becoming of 96. its associate director in 1956, and moved back to Groningen in 1957, where he was “Adriaan Blaauw was one of the most in a key position to contribute to putting influential astronomers of the twentieth the idea of Baade and Oort into practice. century. I had the privilege to be amongst his students when he returned to Leiden ESO was founded in 1962, and Blaauw from his position as Director General. took up the position of Scientific Director He continued to remain keenly interested in 1968 and subsequently became the in everything to do with ESO, and still organisation’s second Director General had his characteristic twinkle in the eye from 1970 until 1974. During that time, when he visited La Silla and Paranal several telescopes, including the ESO ­earlier this year. It is hard to grasp that he 0.5-metre and 1-metre Schmidt tele­ is no longer with us,” said ESO’s cur- scopes, began operating at ESO’s first rent Director General, Tim de Zeeuw. observatory site, La Silla, in Chile. This was also a key period for the design and Adriaan Blaauw was born in Amsterdam, construction of the ESO 3.6-metre tele­ the Netherlands, in 1914. He studied scope, which had its first light in 1976. astronomy at Leiden University, under de Blaauw decided that it was crucial for this Sitter, Hertzsprung and Oort, and project to move ESO Headquarters obtained his doctorate with van Rhijn at from Hamburg to Geneva, to benefit from the Kapteyn Laboratory in Groningen the presence of the CERN engineering in 1946, with a thesis entitled, “A study of group. Today, the 3.6-metre telescope Blaauw was well known for his legendary the Scorpio–Centaurus cluster”. During remains at the forefront of research, host­ patience and wisdom, and for his genu­ his career, Blaauw studied the properties ing the HARPS spectrograph — the ine interest in astronomy and astrono­ of OB associations (groups of young, world’s foremost exoplanet hunter. mers, including the most junior students. hot stars) which contain the fossil record He liked to bring order quietly to most of their star formation history. Perhaps ”The early years for any organisation, par­ topics that he turned his attention to. This his most famous work explained why ticularly one in which half a dozen coun­ included the archives of ESO and of the some of these stars are found to be trav­ tries and their governments are involved, IAU — work which resulted in two books, elling unusually rapidly — so-called run- can provide many challenges. But with ESO’s Early History and a History of away stars. Blaauw proposed that these Adriaan Blaauw’s leadership, preceded the IAU. His personal account of his life stars had originally been members of by Otto Heckmann, ESO started off on a entitled My Cruise Through the World binary systems, but that when one star in strong footing,” said de Zeeuw. of Astronomy, published in the 2004 the binary experiences a supernova ex­ Annual Reviews of Astronomy and Astro- plosion, its companion suddenly ceases After stepping down as Director General physics, provides an extraordinarily to feel the gravitational pull that keeps of ESO, Blaauw returned to Leiden, ­accurate picture of a truly remarkable it in its orbit and hence it “runs away” at and continued to play a very important person, who influenced the lives of many its orbital velocity. This work was pub­ role in international astronomy. He others in a very positive way. lished in his 1961 paper, “On the origin of was President of the International Astro­ the O- and B-type stars with high veloci­ nomical Union from 1976 to 1979, during Tributes to Adriaan Blaauw will appear in ties (the ‘run-away’ stars), and some re­­ which period he managed to convince the next issue of The Messenger. lated problems”. China to rejoin the IAU. Blaauw retired in 1981 and moved back to Groningen, Based on the text of the ESO Announce­ In addition to his distinguished research but stayed active. He served as Chairman ment ann1090. career, Blaauw played a central role in of the Scientific Evaluation Committee setting up ESO. In 1953, the astronomers for the European Space Agency satellite Walter Baade and Jan Oort proposed Hipparcos, advising on many aspects of the idea of pooling European resources its scientific programme. and funding to create an astronomical

The Messenger 142 – December 2010 51 ESO, the European Southern Observa- Contents tory, is the foremost intergovernmental astronomy organisation in Europe. It The Organisation is supported by 14 countries: Austria, G. Filippi – Enabling Virtual Access to Latin-American Belgium, the Czech Republic, Denmark, Southern Observatories 2 France, Finland, Germany, Italy, the Netherlands, Portugal, Spain, Sweden, Telescopes and Instrumentation Switzerland and the United Kingdom. G. Witzel et al. – On the Instrumental Polarisation of NAOS–CONICA 5 ESO’s programme is focused on the P. Hammersley et al. – Upgrading VIMOS 8 design, construction and operation of R. Arsenault et al. – Progress on the VLT Adaptive Optics Facility 12 powerful ground-based observing L. Testi et al. – ALMA Status and Progress towards Early Science 17 ­facilities. ESO operates three observa- tories in Chile: at La Silla, at Paranal, Astronomical Science site of the Very Large Telescope, and at A. Seifahrt et al. – Precise Modelling of Telluric Features in Llano de Chajnantor. ESO is the Euro- Astronomical Spectra 21 pean partner in the Atacama Large Mil- M. Sterzik et al. – Astronomy Meets Biology: limeter/submillimeter Array (ALMA) EFOSC2 and the Chirality of Life 25 under construction at Chajnantor. Cur- R. Gratton et al. – Observations of Multiple Stellar Populations in rently ESO is engaged in the design of Globular Clusters with FLAMES at the VLT 28 the 42-metre European Extremely Large S. Clark et al. – Dissecting the Galactic Super Star Cluster Telescope. Westerlund 1 — A Laboratory for Stellar Evolution 31 R. Maiolino et al. – AMAZE and LSD: Metallicity and Dynamical Evolution The Messenger is published, in hard- of Galaxies in the Early Universe 36 copy and electronic form, four times a year: in March, June, September and Astronomical News December. ESO produces and distrib- J. Walsh – Raymond Wilson Honoured with Two Prestigious Prizes 41 utes a wide variety of media ­connected G. Lo Curto et al. – Availability of Reduction Software for HARPS Data to its activities. For further information, at ESO Headquarters in Garching 42 including postal subscription to The O. Sandu, L. L. Christensen – ESO Participation at the Joint European and Messenger, contact the ESO education National Astronomy Meeting in Lisbon, Portugal 42 and Public Outreach Department at the M. West – Visiting ESO’s Office in Santiago 44 following address: H. Boffin et al. – Café & Kosmos Events in Munich 44 New Staff at ESO – J.-P. Berger 45 ESO Headquarters Fellows at ESO – M. Moerchen, D. Krajnović 47 Karl-Schwarzschild-Straße 2 T. de Zeeuw – In Memoriam Christine Nieuwenkamp 48 85748 Garching bei München Announcement of the “ALMA Community Days: Towards Early Science” 49 Germany Announcement of the ESO Workshop “Fornax, Virgo, Coma et al.: Phone +498932006-0 Stellar Systems in Nearby High Density Environments” 49 [email protected] Personnel Movements 50 www.eso.org In Memoriam Adriaan Blaauw 51

The Messenger: Editor: Jeremy R. Walsh Design: Jutta ­Boxheimer; Layout, Type- setting: Mafalda Martins and Jutta ­Boxheimer; Graphics: Roberto Duque www.eso.org/messenger/

Printed by Peschke Druck Schatzbogen 35, 81805 München Germany

Unless otherwise indicated, all images in The Messenger are courtesy of ESO, except authored contributions which Front cover: Colour image of the barred spiral galaxy NGC 1365 (type SBb) in the are courtesy of the respective authors. near-infrared, taken with HAWK-I on the VLT. NGC 1365 is at a distance of about 18 megaparsecs and hosts an active galactic nucleus. The picture was created © ESO 2010 from exposures through Y, J, H and K filters with exposure times of 4, 4, 7 and ISSN 0722-6691 12 minutes respectively. See Photo Release eso1038 for more details.