The Messenger

No. 135 – March 2009 Headquarters extension building Correlator The ALMA VLTI stars with Wolf-Rayet Galaxy z ~ 2 massat assembly The Organisation

An Extension for ESO Headquarters

Robert Fischer1 who also designed the neighbouring Institute for Plasma Physics in a building Jeremy Walsh1 Max-Planck Institute for Astrophysics. It located about four minutes walk away had 120 offices, an auditorium, work- from the Headquarters. In 2007 more shops, laboratories and storage rooms. offices were rented from the Max-Planck 1 ESO The generous curves and lack of Institute to house the ALMA staff, whose 90-degree corners are two of its defining numbers were increasing rapidly during characteristics. The original building the construction phase. By the end of The ESO Headquarters was completed had four floors and some basement area. 2008 the ESO “campus” consisted of the in 1980, but is now too small to house By the end of the 1980s, however, with Headquarters building, two temporary all the ESO staff and currently only about the gradual increase in numbers of ESO buildings and two locations of rented 50% reside in the original building. staff as the number of telescopes at offices with a staff complement of 444, A decision was taken to seek an exten- the La Silla Observatory increased and only 230 of whom were actually located sion to the Headquarters building in with the development of the Very Large in the original Headquarters building. close proximity to the current one and Telescope (VLT) project, the addition of a a competition was launched for archi­ fifth floor was necessary. When the VLT The difficulties of working closely on joint tectural designs. Three designs were became operational in 1998 and as more projects with offices spread over an shortlisted and the process of selection ESO member-state countries joined, ex­tended area were not just perceived. In for the final design is described. Con- more staff were recruited; the need for 2007 a decision was made to try to im­­ struction will begin in 2010 and is due more offices had to be satisfied by a pre- prove the scientific atmosphere by bring- for completion in 2012. fabricated office block (“portakabin”) in ing all faculty astronomers, fellows, stu- the car park. As more VLT Unit Telescopes dents and science visitors together in the came online, new instruments were main building. Logistically this proved When ESO moved from temporary offices commissioned and the ALMA project en­­ possible only for a short while and new at CERN, Geneva, in 1980 there were a tered its planning phases, staff numbers total of 40 staff members. The Headquar- increased again and a second temporary Figure 1. Montage of the 20 entries for the Architec- ters building was designed by the archi- building had to be erected. In addition, tural Competition for an extension to the ESO Head- tects Hermann Fehling and Daniel Gogel, offices were rented from the Max-Planck quarters.

2 The Messenger 135 – March 2009 scientific staff, e.g., for ALMA, are increas- received and all are shown in Figure 1. The Figure 2. The three prize-winning architectural con- ingly located in offices in the ALMA jury assessed the designs, and awarded cepts for the Headquarters extension building in their revised form. Left: Koch and Partner, Germany; building. The situation is exacerbated by two first prizes (€25 000 each), a third Centre: RISCO, Portugal; Right: Auer + Weber, the small number of meeting rooms and a fourth prize (€15 000 and 10 000 Germany. and the limited size of the auditorium, respectively) in October 2007. One addi- which can only seat about 100. The tional very innovative design, which did (see the aerial photograph on p. 26 of instrument assembly hall and laboratories not however meet the local building regu- The Messenger 130) and the original are also too small for the level of develop- lations, was recognised with a “purchase” com­petition had specified this position. ment and the typical size of second prize (€5000). This land is owned by a consortium generation VLT instruments. An important of local farmers. However negotiations issue was the need for further expansion The top three prize-winning concepts, could not be concluded by the time of the after the European Extremely Large shown in Figure 2, did not fulfill all the revised tender and it was decided to Telescope (E-ELT) design study was ap­­ re­quirements and further technical clarifi- site the building to the southeast on land proved in 2007 and a substantial increase cation was required, such as the link to owned by the Max Planck Society. in scientific and engineering staff is re­­ the current Headquarters building and the The rent for this land will be paid by the quired to develop Phase B (see Spyromilio realisation of the instrument assembly German Federal Ministry of Education et al., 2008). In order to improve com­ hall/laboratory complex (whether as a and Research, just as is done for the land munication and the proximity of the sci- separate building or integrated into the on which the current building is sited. ence, operations, engineering, software extension). The baseline proposal would and administration groups working only have been sufficient to provide towards common goals, to raise the team 476 work places together with the current Final selection focus and to provide ESO with a hallmark Headquarters building. Given the building that embodies its profile, an expected increase to 520 by the time of In the final round of the selection process, increase of office, conference and work- completion, an expansion of 1100 m2 was the three prize-winning architects could shop space by an extension to the considered necessary. The new office revise their original concepts to take into present Headquarters building was pro- and conference extension would then pro- account the decision to have two build- posed. Provision of an extension to vide a total floor area of 9500 m2 and ings as part of the extension — separate the existing building was confirmed by the the technical building 2900 m2. For com- office and workshop/integration build- ESO Council. parison the existing building has a floor ings. The three concepts were assessed area of 10 200 m2 and the storage hall a between June and October 2008. On surface area of 770 m2. Thus the exten- 28 October the ESO Contract Award Architectural designs sion almost doubles the existing area of Committee agreed that the design by the current Headquartes building. By the Auer + Weber best met all the high level The specifications for an extension to time the new building is completed, the requirements by providing, in particular, the ESO Headquarters building were de­­ current one will be more than 30 years old the very important parameter of one fined with careful consideration of future and is in need of major maintenance and signature ESO Headquarters building. expansion and enhancement of the upgrading; it was decided to include this This recommendation of the committee current facilities and a Call for Tender for upgrade in the revised concept. This con- was approved by the ESO Council on architectural concepts was drawn up. cept was then submitted for revised ten- 2–3 December. The same architect’s This was sent to 24 architectural offices der to the three prize-winning architectural office was also responsible for the award- in the ESO member states on 11 July firms with a closing date of 13 June 2008 winning design of the Paranal Residencia. 2007 with a deadline for entries of 24 Sep- with the aim of achieving “one signature tember. An international jury consisting headquarters”. A fixed fee of €75 000 was of eminent architects, representatives of awarded for each submission. The new extension the town of Garching and ESO was constituted to judge the designs. Twenty Initially it was planned to build the exten- The first concept submitted by Auer + responses to the Competition were sion to the south of the current building Weber was already marked out by its

The Messenger 135 – March 2009 3 The Organisation Fischer R., Walsh J., An Extension for ESO Headquarters

laboratory is a circular building and, although separate, will be accessible by a tunnel from the existing building. A large cleanroom will be housed within the workshop building.

As part of the process of approval for the building plans, a presentation to the local authorities took place at ESO Garching in December 2008. The ESO Director General Tim de Zeeuw outlined the rea- sons for seeking a new building and the proposed architectural solution was presented by Professor Fritz Auer from the office of Auer + Weber. The planned extension received the full support of the town of Garching and the District Office of Munich.

The final aspects of the design are cur- rently being agreed and construction work is expected to begin in early 2010, with the new extension ready for occupancy in 2012. With the ESO flag flying above its signature Headquarters, the staff will be in pole position to manage the current activi- ties — La Silla Paranal Observatory, ALMA, construction of the E-ELT — and Figure 3. Ground plan of the Auer + Weber design for further in the dimensions of several to forge the next large astronomy project. the ESO Headquarters. The single circular building to aspects, such as internal light wells that the north is the technical building. are identical to those of the ESO tele- scope mirrors (2.2, 3.6 and 8.2 metres). References harmony with the existing building; the The two buildings will be connected Spyromilio, J. et al., The Messenger, 133, 2 revised concept further emphasises the by an enclosed bridge whose length is on impression of community between the the natural axis of the current building two parts. There is, in close reference to (see Figure 3). Figure 4 shows an artist’s Figure 4. Side view of the Auer + Weber design the original design by Fehling and Gogel, impression of the new building from the for the ESO Headquarters extension from the extensive use of circles and this is taken southwest. The workshop and integration southwest.

4 The Messenger 135 – March 2009 Credit: ESO/A. Wicenec ALMA Operations Site (AOS) Technical Building. Technical (AOS) Site Operations ALMA the at place in shown is quadrant first the Here Correlator. ALMA the — site astronomical an at used ever computer operations/second) (17 peta- fastest m) and (5000 highest The

Telescopes and Instrumentation Telescopes and Instrumentation

The ALMA Correlator: Performance and Science Impact in the Millimetre/Submillimetre

Alain Baudry1 2 domain of the cross-correlation functions Specifications and architecture (or the interferometer complex visibilities). Moreover, time offsets or time lags are The hardware sub-systems, in which the 1 University of Bordeaux, LAB, France easily implemented in digital systems and digital filtering and correlation functions 2 European ALMA Project Office, ESO thus the spectral information contained have been implemented for the ALMA in the observed sources is also available main array and the associated firmware, from a digital correlator; the correlation lag form what is referred to as the ALMA The basic properties of digital corre­ functions and the astronomical spectra baseline correlator system (see specifica- lation are introduced and the main are related by a time/frequency Fourier tions in Table 1). blocks that form the correlator system transform. designed for the ALMA main array are The ALMA correlator designers have described. Some technical challenges There are two main correlator architec- faced two major challenges, driven by and the performance of this system tures: (a) the XF architecture where the main science goals in the millimetre are presented, together with examples the X-part (cross- or auto-correlation part) and submillimetre: (a) process a broad of observational modes, total band- is performed first and the Fourier trans- signal bandwidth of 16 GHz (8 GHz per widths and spectral resolutions. The form (F-part) is applied at a later stage to polarisation) from each of 64 antennas high flexibility of the ALMA correlator is analyse the source spectral properties; (each antenna provides 96 Gbits/s); emphasised and its ability to bring (b) the FX architecture where the Fourier (b) extract from the input band various new data in molecular line or continuum transform is performed first. In terms spectral windows and provide flexible astrophysics projects is discussed. of signal processing the XF or FX spectral resolutions. Figure 1, adapted ap­proaches are equivalent. ALMA has from Escoffier et al. (2005), presents adopted two different types of correlator. the baseline correlator system schemati- ALMA signal processing and correlation One FX correlator constructed by a cally. This includes “correlator station Jap­anese team processes the signals col- electronics” boards processing the digi- The Atacama Large Millimeter/submillim- lected by 16 antennas of the ACA. The tised data streams from each antenna eter Array (ALMA), under construction second correlator, constructed by an in the array and “correlator baseline elec- at the 5000-metre high altiplano of north- NRAO/European “integrated team”, proc- tronics” boards providing the interfero- ern Chile, will explore the Universe in esses up to 64 ALMA antennas. It is a metric cross-products from up to 2016 the millimetre/submillimetre wave range highly flexible “Digital hybrid XF” design independent antenna pair combinations up to 1 THz (300 micrometres). With in which the 2 GHz input baseband (for 64 antennas). an impressive collecting area of 50 + 16 is digitally split into several sub-bands to large antennas moveable over an area enhance the spectral resolution, a con- We briefly describe below some of the from about 150 m to more than 15 km in cept first proposed in the European Sec- main parts of the ALMA correlator and diameter, ALMA offers an unprecedented ond Generation Correlator (2GC) study. refer to the main blocks in Figure 1: angular resolution and imaging capability The Digital Hybrid XF correlator now (see Haupt & Rykaczewski, 2007). The merges the main ideas from the initial Filtering: After digitisation and time 50+ (up to 64) 12-metre antennas form NRAO parallel array XF architecture with demultiplexing of the antenna signals for the ALMA main array; four 12-metre and the frequency division scheme of the each baseband, frequency division is twelve 7-metre antennas form the ALMA 2GC study. More generally, it is important accomplished in the Tunable Filter Bank Compact Array (ACA). The radio waves, to stress that digital filtering/processing (TFB) boards placed in the station once they have been captured with the offers great advantages with respect electronics racks. These boards divide ALMA antennas, are converted into to analogue signal processing: reproduc­ the 2 GHz baseband into 32 frequency- low frequency signals (while preserving ibility of performance, stability with mobile sub-bands of 62.5 MHz. Each the phase information of the incoming respect to temperature drifts, easier cali- sub-band is independently processed, waves), digitised in specific modules and bration and higher flexibilty. assigned to one of the 32 correlator finally combined in a large digital ma­­ chine, named the correlator. The digital correlator is the system combining the Item Specification outputs of all antennas, as selected in the Antennas Up to 64 (up to 2016 interferometric baselines) array, to detect the astronomical source Baseband inputs per antenna 8 x 2 GHz power by measuring the cross-correlation Input sample format 3-bit, 8-level at 4 Gsample/s coefficients of all antenna pairs in addi- Output correlation sample format 2-bit, 4-level or 4-bit, 16-level tion to the auto-correlation coefficients of Processing rate 125 MHz each antenna. From these coefficients, Spectral points per baseband (Frequency Division Mode) Up to 8192 per correlator quadrant and by further processing of the data, Spectral points per baseband (Time Division Mode) 64, 128 or 256 per correlator quadrant images of the astronomical sources are Polarisation products 1, 2 or 4 obtained. In mathematical terms the source image is directly related to the Table 1. Main top level specifications of the ALMA Fourier transform in the spatial frequency baseline correlator

6 The Messenger 135 – March 2009 AT ANTENNA CORRELATOR STATION ELECTRONICS ONE BASEBAND POLARISATION PAIR

POL 1 4 GHz DIGITIZER TUNABLE FILTER BANK CARD DATA DATA 32 - PARALLEL OUTPUTS × 2 - BITS STATION CARD TRANSMITTER RECEIVER PER SAMPLE × 2 - POLARISATIONS POL 2 4 GHz DIGITIZER TUNABLE FILTER BANK CARD ER ER 4 GHz DIGITIZER TT TUNABLE FILTER BANK CARD DATA DATA LI STATION CARD TRANSMITTER RECEIVER SP

4 GHz DIGITIZER COMBIN TUNABLE FILTER BANK CARD L L CA CA 4 GHz DIGITIZER TUNABLE FILTER BANK CARD

DATA OPTI DATA

OPTI

STATION CARD TRANSMITTER RECEIVER 4 GHz DIGITIZER 12 TUNABLE FILTER BANK CARD 1: :1 12

4 GHz DIGITIZER TUNABLE FILTER BANK CARD DATA DATA STATION CARD TRANSMITTER RECEIVER 4 GHz DIGITIZER TUNABLE FILTER BANK CARD

DIGITIZES ANALOGUE SIGNALS AND THREE OPTIC THREE OPTIC FINE DELAY AND BULK DELAY, RE-ORDERS AND APPLIES ULTRA FINE DELAY CARRIERS CARRIERS BANDWIDTH SELECTION PACKETIZES SAMPLES

CORRELATOR BASELINE ELECTRONICS

CORRELATOR ARRAY (1 OF 4 IN SYSTEM SHOWN)

32 - PARALLEL OUTPUTS × 2 - BITS PER SAMPLE × 2 - POLARISATIONS ANTENNA 63 63 31

E LONG TERM COMPUTER AN E TO 2 ACCUMULATOR INTERFACE TRIX E PL AN 1 MA E PL

DRIV AN 0 L E PL OR LTA ADDER REAL TIME AN TA AT

PL TREE COMPUTER EL 32 - PARALLEL OUTPUTS × 2 - BITS PER SAMPLE × 2 - POLARISATIONS HORIZ ON CORR ANTENNA 0 0 4 PCB 63 0

VERTICAL DRIVE TO CORRELATOR MATRIX

“planes” (see below) and the resulting signal is multiplied by its time- (or lag-) Figure 1. Block diagram of the ALMA 64-antenna spectra are later stitched together to shifted version to derive the correlation correlator showing the main components of the cor- relator antenna and correlator baseline electronics form a final spectrum with 32 times more coefficients. Each basic 256-lag circuit in (adapted from Escoffier et al., 2005). The digitiser spectral channels across the original a single correlator chip is driven by the and data transmitter boards at the antennas are also baseband. Distributing the correlator two polarisation signals sent from each shown. The digitised data flow is carried to the cor- plane resources to fewer than 32 digital receiver band and there is the further relator system by optical fibres. filters narrows the input bandwidth ability to address fewer lag blocks in the and increases the spectral resolution. The chip to support double polarisation digitiser also in­­cludes a time-demultiplexed 62.5 MHz sub-band is extracted from modes or a full Stokes parameter analy- stage consistent with the 250/125 MHz 2 GHz in a three-stage digital filter struc- sis of the incoming waves. clock rates. The correlator system is ture implemented in a programmable a synchronous machine with a very large logic device (or Field Programmable Gate Two other system aspects are critical: number of filter and correlator boards Array, FPGA). The last stage determines signal digitisation and data transmission (1024) in many racks that makes data the final filter sub-band characteristics from station electronics to correlator trans­mission a difficult problem. The and enables the bandwidth to be further board electronics. Prior to digital correla- processing clock rate in all racks is at narrowed to 31.25 MHz by downloading tion, the analogue baseband signals 125 MHz, but the station rack to correla- a specific set of pre-calculated digital are converted to digital samples in a digi- tor rack communication system involves weights. tiser module specifically designed for 16 384 cables working at 250 MHz. The ALMA. These modules, plugged into the output phases of these cables are re­­ Correlation: All independent cross- antenna digital racks, are not part of motely controlled and adjusted at the link products for all 64 antennas are derived the correlator system, but their reliability frequency for error-free data transmission. from the 32 correlator planes of the and efficiency are essential in the signal baseline electronics boards (see Correla- processing chain. The data encoding The correlator system is basically tor Array in Figure 1, lower). A correlator is made on 3 bits and corresponds to a or­ganised by quadrants, each quadrant plane is a 64 x 64 matrix formed by theoretical 96% cor­relation efficiency. processing one baseband pair, to four printed circuit boards, each with sev- Because the ALMA correlator cannot accommodate four baseband pairs for eral assembled specific integrated process the data at the 4 GHz sampling two polarisations per antenna. This circuits (the correlator chips) in which the rate (for 2 GHz input baseband), the architecture makes it possible to enhance

The Messenger 135 – March 2009 7 Telescopes and Instrumentation Baudry A., The ALMA Correlator

Figure 2. Schematic layout of one quadrant of the DIGITISED DATA STREAMS FROM ANTENNAS AND FIBRES correlator system. Station and correlator baseline printed circuit boards are distributed in bins and racks and the correlator outputs are passed to the STAT ION ELECTRONICS CORRELATOR BASELINE STAT ION ELECTRONICS RACKS ELECTRONICS RACKS RACKS long-term accumulation boards and the real-time correlator data processing computers.

Bin 1 Bin 1 250 MHz Bin 1 Bin 1 Bin 1 Bin 1 250 MHz Bin 1 Bin 1 Bin 2 Bin 2 rack-to-rack Bin 2 Bin 2 Bin 2 Bin 2 rack-to-rack Bin 2 Bin 2 Bin 3 Bin 3 and Bin 3 Bin 3 Bin 3 Bin 3 and Bin 3 Bin 3 Bin 4 Bin 4 self-test link Bin 4 Bin 4 Bin 4 Bin 4 self-test link Bin 4 Bin 4 20 bits of integration and secondary storage for readout of the correlation coefficients. The 32 768 chips required for full operation of the four quadrants have been industrially produced. Use of these CORRELATOR DATA PROCESSING COMPUTERS chips in the correlator environment of the integration centre and of the ALMA high site has proved to be extremely reli- the total bandwidth, if that were to the industrial production phase of the able. The correlator chip resources can be required, by adding another quadrant. TFB cards. These included questions as also be combined to provide 4-bit x 4-bit Figure 2 shows the schematic layout diverse as the Ni–Au finishing of the multi­plication so that the quantisation of a single correlator quadrant. There are cards for long term protection or defini- efficiency is enhanced, thus improving 32 racks in total (8 racks per quadrant) tion of the optimum temperature profile ALMA’s interferometric sensitivity overall. with, in addition, power supply racks and for reliable assembly of components the Correlator Data Processor and Corre- free of lead (to meet the EC directive on The correlator system is a very large lator Control computer racks. All racks Restriction Of Hazardous Substances). specific computing machine. With 4096 and computers are installed in the correla- Special test equipment and test proce- multipliers per chip performing 2-bit tor room of the 5000-metre altitude Array dures were developed to validate the correlation at a 125 MHz clock rate and Operations Site (AOS) technical building digital filter functionality and check the for 64 chips per correlator board, the (see photograph on p. 5). industrial production quality. This testing number of operations performed per sec- was required because of the complexity ond in the full system is 1.7 x 1016. The of the printed circuit board, the large computing power in a single TFB board Technical challenges and performance number of ball grid array connections to is already impressive. With several hun- each board for each large FPGA and the dreds of multi-bit (typically 8 to 16) adder Novel hardware or firmware develop- large number of TFB boards to be pro- and multiplier stages implemented in ments have been made to meet the ALMA duced (512 for 8 basebands per antenna one digital filter, around 1012 operations baseline correlator specifications. Two x 64 antennas, see Figure 3). per second are performed in a board. examples are given for illustration: The correlator output streams must reach the digital filter subsystem conceived by The ALMA correlator chip is a specific, a reasonable rate for final processing the European team; and the correlator highly integrated microcircuit providing and archiving. From the basic accumula- chip designed by NRAO. Several original an unsurpassed number of multipliers in tion mode of the ALMA correlator chip, aspects of the TFB design concern the a single custom-made chip. There are the Long Term Accummulator (LTA in ability to move sub-bands and implemen- 4096 multipliers (also named “lags”) per Figure 1) transfers 16, or integer multiples tation of the firmware required to meet chip providing 2-bit x 2-bit multiplication, of 16, milliseconds of integrated data to the challenging filter specifications (e.g., high stop band rejection, low baseband ripple and low power dissipation). All filter Figure 3. A small series of boards required functions must be implemented without for digital filtering before correlation. 512 boards have been produced and function- exhausting the resources available in high ally tested for the full four-quadrant sys- performance FPGAs purchased from tem. industry. We have selected a large FPGA with recent 90-nanometre technology to implement two 62.5 MHz sub-band filters in one FPGA. A matrix of 4 x 4 FPGAs is required for frequency division of each baseband and these 16 devices are assembled on a multi-layer TFB board together with other components (fine interferometric delay tracking is imple- mented in other small FPGAs). Several difficulties were met and solved during

8 The Messenger 135 – March 2009 the “real-time computer system” to Figure 4. Front panel image of two process the correlation coefficients fur- bins in a Station Electronics rack showing TFB cards, station cards and ther. The maximum correlator output other system cards (see Correlator capacity reaches 1 GByte/s (in 16 data Station Electronics in Figure 1). The streams) or 256 MByte/s per correlator pairs of holes seen in front of each quadrant (in four data streams). Lowering TFB card allows cool air from the correlator room to be blown into the the output rate is equivalent to sacrificing rack. interferometric baselines.

Because the ALMA baseline correlator is such a large machine, power dissipation is a difficult task at the high altitude AOS, where the air density is roughly half that at sea level. There was an initial concern with the TFB boards because they im­­ plement several functions (and thus dissi- pate much power) and are plugged in racks where the air circulation is slowed down for practical reasons. With 75 W dissipated power per TFB board, the first engineering goal of 100 W was well met. The latest firmware optimisation and the develoment of a newer filter design have further reduced the dissipation below 60 W. This new performance low- ers the temperatures inside the FPGAs, leading to a longer lifetime of these com- ponents. Based on the power dissipation measured at the high site with the first The first quadrant of the 64-antenna cor- specifications are similar (except that the quadrant, we anticipate a total power dis- relator has been shipped to Chile, installed ACA correlator processes 16 instead sipation of 150 kW for the four-quadrant in the AOS technical building and was of 64 antennas). In the Japanese FX-type system including the real-time data proc- successfully commissioned in the summer correlator, spectral averaging is per- essor and control computers. of 2008 (Figure 4). The air circulation formed after the F-part and the correlation from the floor up to the top of the station is directly performed in 4 bits, but these racks and temperature throughout the details should remain invisible to the First operational results and installation bins and racks are remotely controlled. astronomers; only the effective spectral at AOS The 64-antenna correlator room is next to resolution should differ slightly. the antenna “patch panel” room, which In parallel with the construction of the enables the fibre outputs of each antenna large 64-antenna correlator, two scaled- pad to be physically connected to the Correlator modes for astronomical down models of the large machine have correlator inputs. The patch panel sup- science been made. One for the ALMA Test ports connections to about 200 antenna Facility (ATF) consisting of two prototype pads for all moveable 50 + 16 antennas The ALMA baseline correlator supports antennas at the NRAO VLA site (Very and allows the ACA antenna data to a broad variety of observing modes, Large Array, New Mexico) and one for the be processed in the ALMA baseline cor- which may be classified into two main 3000-metre high ALMA Operations Sup- relator. When all 50 antennas of the main categories, the Time and Frequency port Facility (OSF) site. The first scaled- array are combined in the 64-antenna Division Modes (TDM and FDM). In the down model was moved to the ATF site correlator with 14 of the 16 ACA antennas TDM operation mode there are 32 parallel in 2007 and used through to the end of we then reach the ultimate ALMA sensi- correlator planes, each processing, at 2008. First successful end-to-end interfer- tivity. For compact sources the sensitivity 125 MHz, one-millisecond signal time ometric operation was demonstrated on may be improved by about 8% with slices from the digitised input baseband, the sky in January 2008 in the direction of re­spect to the main array alone and all and the correlator behaves as a pure the Orion Nebula (see Laing, 2008) where calibration sessions are performed more XF system. In the FDM operation mode many spectral lines were observed at the quickly. The ACA 16-antenna correlator each of the 32 sub-bands in the 2 GHz expected frequencies. The second two- has also been installed on the high site. input baseband is processed in one antenna correlator was installed at the Processing the same data (up to 16 anten- of the 32 correlator planes, or fewer sub- OSF in 2008. It is used to test the produc- nas) in two different correlators is of bands are processed by all planes to tion antennas and the associated equip- great value for an advanced comparison enhance the spectral resolution. The TFB ment before they are moved to the AOS. of these two large machines, whose filter boards are designed to support both

The Messenger 135 – March 2009 9 Telescopes and Instrumentation Baudry A., The ALMA Correlator

Total Bandwidth Resolution (Nyquist, kHz) Resolution (Double Nyquist, kHz) Table 2. Bandwidth and resolution for 2-bit correla- 2 GHz 244 488 tion with one 2 GHz baseband processed per corre- lator quadrant in frequency division operation mode 1 122 244 (FDM). 500 MHz 61 122 250 30.5 61 § Available with specific digital weights downloaded 125 15.3 30.5 in last stage of the digital filter. 62.5 7.6 15.3 31.25§ 3.8§

TDM and FDM, the data being directly with the same spectral resolution, sensi- respectively. They have been arbitrarily sent to the final requantisation stage tivity and polarisation options; (c) multi- selected to illustrate our discussion of each TFB board in the TDM operation spectral resolution over different band- on required bandwidth. The total velocity mode. Examples of the spectral resolution widths to zoom on specific spectral coverage corresponds to the expected achieved in FDM observing modes, for features. All quadrants are independent velocity extent at the base of the line the case of only one baseband processed and these different modes could also profile with some additional spectral noise per quadrant, are given in Table 2 for be implemented simultaneously with over- channels on each side of the line fea­ different input bandwidths. Nyquist sam- lapping quadrants. ture(s) of interest to provide a reference pling is the basic working model of the intensity level. correlator, but double Nyquist sampling is also available to im­­prove the correlation Spectral line and continuum astrophysics The maximum bandwidths that the base- efficiency (7% better) at the expense line correlator can process match rather of lower correlator resources and so lower The ALMA baseline correlator, with its well with the total bandwidths in Table 3. spectral resolution. The maximum reso­ flexible resolution and ability to analyse These maximum bandwidths are: 2 or lution achievable, 3.8 kHz, enables multi-spectral windows, is extremely 1 GHz, 500, 250, 125, 62.5 or 31.25 MHz, extremely detailed spectral studies to be well adapted to any type of spectral work. with one quadrant; 4 GHz with two quad- carried out. Two parameters are of interest, the total rants; or 8 GHz with four quadrants. bandwidth and the spectral resolution: If two basebands are processed per Spectral resolution: The various resolu- quadrant, the resolutions shown in Table 2 Total bandwidth: Table 3 shows the typi- tions available with the baseline correlator are twice as poor. There is another cal total velocity coverage required for are suited to a large variety of astro­ degradation by a factor of two with the line analysis in a number of Galactic and physical environments. Table 2 gives additional polarisation cross-products extragalactic molecular sources, or examples, but there are more selectable required for a full Stokes parameter anal- in planets, and the corresponding total modes with the 4-bit correlation and ysis. In order to improve the correlation bandwidths at two widely separated fre- polarisation cross-products options. We efficiency, 4-bit correlation is also sup- quencies (around 90 GHz where mo­­ give a few examples. About 1 MHz ported, but the frequency resolution lecular transitions from HCO+ or HCN are resolution is convenient to provide many is now four times poorer than that shown observed, and around 602 GHz where details in the CO J = 2–1 line of many in Table 2. Double Nyquist sampling methanol is present). These frequencies nearby galaxies or in energetic galactic is also available with 4-bit correlation; it fall in the ALMA receiver Bands 3 and 9, outflows. This can be achieved, for does not improve the already high effi- ciency much, but decreasing the spectral resolution speeds up the data dump rate. Source Typical Total Velocity Coverage Total Bandwidth About 70 different observing modes (km/s) 90 GHz, 602 GHz are available with only a few TDM modes. Galactic Sources TDM is to be preferred for fast dump Energetic Outflow in Young Stellar Objects 600 180 MHz, 1.2 GHz rates (16 ms minimum) and moderate Spectral Line Survey 300 90 MHz, 600 MHz resolution (31.25, 15.6 or 7.8 kHz) across Orion, Galactic Centre 80–160 24–48 MHz, 161–321 MHz 2 GHz. Compact H ii regions 40 12 MHz, 80 MHz Molecular Cloud Spectra 10–40 3–12 MHz, 20–80 MHz The 64-antenna correlator has the ability Dark Clouds 5 1.5 MHz, 10 MHz to move spectral sub-bands within the Extragalactic Sources input baseband and to split one correlator Nearby Galaxies (≤ 200 Mpc) ≤ 2000 ≤ 0.6 GHz, ≤ 4 GHz quadrant into independent sub-units Highly Redshifted Sources As large as possible so that several options are supported: Planets (a) high resolution in a given spectral region Pressure Broadened Lines 1000–3000 0.3–0.9 GHz, 2–6 GHz from 2 GHz to 62.5 MHz (or 31.25 MHz) with 62.5 MHz sub-bands tunable any- Table 3. Examples of total velocity coverage required where within 2 GHz; (b) multiple disjoint for line observations of Galactic and extragalactic spectral regions fitting within 2 GHz but sources.

10 The Messenger 135 – March 2009 example, with 1 GHz total bandwidth in to astrobiology and the roots of prebiotic confusion may become essential. Despite the Nyquist sampling TFD mode and chemistry in the interstellar gas, the these difficulties we believe that the with the high sensitivity 4-bit option; the range of different environments in our Gal- ALMA correlator will greatly contribute to resolution for the CO J = 2–1 line is axy and distant galaxies, etc. Extremely mo­lecular astronomy in the mm/submm then 1.3 km/s. At the higher frequencies rich spectra and complex molecules, domains because of the following ad­­ of CO, one may either bin the spectral prebiotic or not, have been observed with vantages: (a) in a broad bandwidth many channels or use the coarser resolution of several existing millimetre-wave tele- lines are detectable at once and this the TDM mode (7.8–31 MHz). On the scopes in a broad range of objects (very allows the observer to identify the regions other hand, there are a number of inter- young stellar objects, pre-stellar cores, in crowded line spectra with less confu- esting astrophysical cases where velocity nearby or starburst galaxies, etc.). In the sion; (b) multi-spectral resolution and very resolutions as high as 0.02–0.05 km/s Orion Nebula or in the Galactic Centre, high spectral resolution help to resolve are required: study of protostellar discs, crowded line spectra and “forests” of lines line blends; (c) multi-spectral windows dark molecular clouds, wind velocities in have been observed, so that identifica­- offer the ability to compare properties of planets or thermal line widths in comets. tion of the molecular carrier of a given line various molecules in the same source; In ALMA Band 3 for instance, around becomes difficult. An example of the (d) resolution can be traded for sensitivity 6 kHz resolution is needed to reach 0.02 crowded spectra obtained with the IRAM options or polarisation modes. Of course, km/s and the observing modes providing interferometer is given for Orion-KL, in complex molecular sources, and with 3.8 or 7.6 kHz are thus well suited. Analy- a massive star-forming region, around adequate sensitivity, interferometric maps sis of Zeeman splitting or cosmic maser 223 GHz (see upper panel of Figure 5, showing different spatial distributions lines also requires high resolutions. adapted from Favre et al., 2008). An even may also help in the spectral confusion more extreme case is that of Sgr B2(N) problem (see lower panel in Figure 5 The ALMA baseband correlator will be where Belloche et al. (2008) detected showing good spatial separation of mod- configured for a number of broadband about 100 lines every 1 GHz around erately blended lines). continuum projects in Galactic or extraga- 100 GHz. From 88 transitions free of any lactic sources. The TDM mode is ideal spectral contamination, and from line for processing 2 GHz bandwidths with maps in a subset of these lines, Belloche Figure 5. Upper panel: Cross-correlation spectrum low spectral resolution in single, double or et al. have identified the complex amino of the Orion Nebula showing spectral features of full polarisation observing modes. In all acetonitrile molecule — a potential pre- complex molecules observed with the IRAM interfer- modes, combining independent quad- cursor of glycine. Crowded line spectra ometer around 223 GHz. Lower panel: Spatial distri- rants broadens the total bandwidth if that are expected in several ALMA submillimet- bution of the ethylene glycol and formamide mole- cules obtained with the IRAM interferometer; the red were necessary. Moderately broad band- ric bands and spectroscopic databases circle corresponds to the spatial extent of the width will be sufficient in projects such or new techniques to predict spectral 30-metre telescope beam. as the thermal dust temperature study of Galactic young stellar objects and polari- PdB spectrum sation imaging of dust in young objects 8 At position marked

by * on the maps e e or molecular clouds in order to investigate below ol at 6 yc

the role of the magnetic field. FDM modes gl ) rm

rmamid (K with 0.5–1 GHz bandwiths may often fo T Fo yl

4 lene be appropriate depending on the selected hy eth Et receiver band. There is always the option 2 M to perform continuum and spectral line observations simultaneously with overlap- 0 ping quadrants. For example 2 GHz 2.2348 2.2347 2.2346 2.2345 2.2344 2.2343 continuum with low spectral resolution in Rest Frequency (MHz x 10 5) TDM mode can be combined with higher FDM frequency resolution mode in 2 GHz –5°22´25˝ Ethylene glycol –5°22´25˝ Formamide or lower bandwidth.

–5°22´30˝ IRc2 –5°22´30˝ IRc2 30m beam 30m beam An outstanding tool for molecular * * complexity –5°22´35˝ –5°22´35˝

The ALMA correlator will offer new oppor- –5°22´40˝ –5°22´40˝ tunities to help understand the complexity observed in the molecular Universe. Complexity refers to aspects as different –5°22´45˝ –5º22´45˝ as the identification or discovery of new molecules in the interstellar medium or in 5h35m15.s0 14.s514.s013.s5 5h35m15.s014.s514.s013.s5 circumstellar envelopes, the relationship

The Messenger 135 – March 2009 11 Telescopes and Instrumentation Baudry A., The ALMA Correlator

Future possibilities — near and longer one sub-array will track a source while the the ACA and baseline correlator perform- term second one will perform antenna commi- ance becomes easier. ALMA science will sioning tasks. Each correlator quadrant benefit from these future possibilities. The first quadrant of the baseline cor­ may support more than two sub-arrays. relator system installed in the AOS techni- cal building (see p. 5) supports up to The summed output signals from two or Acknowledgements 16 antennas. It is available for interfero- more antennas in the array will also The ALMA 64-antenna correlator has been con- metric commissioning tasks and it be available in the correlator to form a big structed by a large group of people within the ALMA meets the conditions for early science single antenna which can be combined Correlator Integrated Product Team with support from requiring 16 antennas and a basic set of with other antennas located on other the American and European ALMA Executives and spectral line modes. The first call for continents to provide extremely high spa- from the various institutes involved in the design, pro- duction and testing: in the USA, NRAO, Charlottesville ALMA proposals is ex­­pected around 2011. tial resolution. This observing mode, where the correlator quadrants are first assembled; in One may anticipate that several FDM named Very Long Baseline Interferometry Europe, LAB at Université de Bordeaux, Osservatorio and TDM operational modes including (VLBI), will require some additional equip- di Arcetri, Florence and Astron, Dwingeloo. It is a double or full polarisation modes with dif- ment at the AOS, but will certainly be great pleasure to acknowledge the great spirit of co­­ operation which always prevailed between NRAO, ferent total bandwidths will be scheduled used in the future. Pulsar observations Charlottesville and Université de Bordeaux from the for early science. The second correlator may be supported as well, provided that design reviews to the ultimate construction/testing quadrant will be installed in 2009 and pulsar period models can be implemented phases. The ALMA correlator software group played the third and fourth quadrants, required and that the ALMA receiver bands are of a key role in all the integrated tests of the correlator system. Highly stimulating exchanges developed to support up to 64 antennas, could be interest in pulsar astrophysics. between the European and Japanese correlator assembled by the end of 2010. teams at the time when the European Second Gener- Finally, thanks to the ACA patch panel and ation Correlator concept was emerging. Other correlator configuration modes will the 16-antenna correlator being installed become available in the future. Sub- next to the 64-antenna baseline correlator, References arraying, namely the ability to process the following can be achieved: (a) the over- independent subsets of antennas oper- all ALMA sensitivity can be enhanced Belloche, A. et al. 2008, A&A, 492, 769 ated in different modes, is an important by combining all data from a maximum of Escoffier, R., Webber, J. & Baudry, A. 2005, ALMA System document, ALMA-60.00.00.00-001-B- ALMA feature. Initial implementation 64 antennas; (b) calibration of the twelve SPE of two sub-arrays is a high priority. For smaller 7-metre antennas can be efficiently Favre, C. et al. 2008, The Molecular Universe instance, one group of antennas will map accomplished when they are cross-corre- International Meeting, Arcachon, 5–8 May 2008 a source, while just one other antenna lated with all other 12-metre antennas; Haupt, C. & Rykaczewski H. 2007, The Messenger, 128, 25 will perform single dish observations, or (c) comparison and cross-calibration of Laing, R. 2008, The Messenger, 132, 28

The ALMA Optical Fibre Network Patch Panel, shown left, was successfully installed in the AOS technical building at 5000 m altitude on the Chajnantor Plateau in December 2008. The Patch Panel, an ESO deliverable to ALMA, is used to send the Central Local Oscilla- tor signal to all the ALMA antennas on Chajnantor, to receive the astronomical signals and redirect them to the Ampli- fier and Demultiplexer, before the real correlation takes place in the ALMA Correlator.

12 The Messenger 135 – March 2009 Telescopes and Instrumentation

Improving the Multiplexing of VIMOS MOS Observations for Future Spectroscopic Surveys

Marco Scodeggio1 VIMOS spectrograph (Le Fèvre et al., separate shorter exposures, and the tele- Paolo Franzetti1 2002) is its high degree of multiplexing, scope is offset by a small amount after Bianca Garilli1 conceived specifically to speed up each exposure (typically 1 or 2 arcsec- Olivier Le Fèvre2 the execution of spectroscopic surveys onds), making sure that the objects are Luigi Guzzo3 significantly. Unfortunately, the VIMOS kept inside the MOS mask slits. As a CCDs installed when the instrument was result of the offsets, the object spectra commissioned are thinned E2V detectors fall on different pixels on the CCD in 1 INAF IASF-Milano, Italy from early 2000 technology, and they are different exposures, and it is possible to 2 LAM, Marseille, France affected by significant fringing redwards obtain a relatively accurate and complete 3 INAF Osservatorio Astronomico di of approximately 800 nm, as shown in reconstruction of the fringing pattern by Brera, Merate, Italy Figure 1. Without proper corrections the median-combining all the available expo- spectra obtained with VIMOS red grisms sures. (including the LR_red, MR, HR_orange The need to reduce the negative effects and HR_red grisms) for faint extragalactic The cost of implementing this observing of fringing on VIMOS spectra has sources are very difficult to use above technique is a reduced multiplexing led astronomers to use observing tech- this wavelength. The effects of fringing for MOS observations. To make sure the niques that significantly limit the need to be counteracted to obtain spec- target objects remain visible inside multiplexing of VIMOS observations in tra that can take full advantage of the the slits after the offsets, the slits must be multi-object spectroscopy mode. In wavelength coverage provided by these designed and cut longer than they would this paper we propose a new observing grisms (reaching approximately 950 nm), otherwise need to be (in practice, one strategy which, coupled with a new data and in order to extend the wavelength has to specify a larger sky region inside reduction technique, has the potential and the redshift coverage of the redshift the VIMOS mask preparation software to double VIMOS multiplexing while pro- surveys as much as possible. This is [VMMPS]). In stare mode, with a typical ducing spectra of a quality comparable why the data for the two main surveys faint object size of 2 arcseconds (the to that obtained in the major surveys carried out so far with this instrument in average apparent diameter of a high red- performed so far. MOS mode, the VLT VIMOS Deep Survey shift galaxy in ground seeing conditions), (VVDS; Le Fèvre et al., 2005), and the and a minimum sky region on each side zCOSMOS survey (Lilly, 2008), and the of the object of 2 arcseconds to allow Modern spectroscopic surveys are gener- data for many other smaller VIMOS pro- for an accurate sky subtraction, slits ally based on multi-object observations, grammes, have all been obtained starting would typically be 6 arcseconds long (i.e., where many tens to hundreds of objects from observations carried out in jitter 30 VIMOS CCD pixels). In jitter mode, are observed simultaneously during each mode. The total exposure time for each to accommodate a pattern with five jitter telescope pointing, to speed up the com- instrument pointing is subdivided into positions (like those used for VVDS pletion of projects that collect the spectra and zCOSMOS observations), we must of many thousands of objects. As it is add another 2 to 3 arcseconds on often the case with astronomical instru- each side of the object, for a total slit mentation, the practical requirements length of 10 to 12 arcseconds (i.e., 50 to to achieve this high degree of multiplexing 60 VIMOS CCD pixels). This approximate are quite at odds with those necessary doubling of the typical slit length directly to obtain high quality spectra for each of translates into a reduction of 50% of the surveyed objects. In this paper we the VIMOS multiplexing, which is precisely describe an observing strategy and a what has happened for both the VVDS specific data reduction procedure for the and zCOSMOS projects. ESO Very Large Telescope (VLT) VIsible Multi-Object Spectrograph (VIMOS), designed to increase the multiplexing of Searching for alternatives MOS observations, without significantly affecting the quality of the spectra thus In preparation for future spectroscopic produced. In the optimal case of a deep surveys, we have recently studied the survey with a large set of potential targets possibility of adopting a different observ- for the spectroscopic observations, it ing strategy to increase the multiplexing will be possible to double the multiplexing of VIMOS observations in MOS mode, of the MOS observations with VIMOS without significantly affecting the data using this new strategy. quality and the measurement reliability of redshifts. As a first step we started Figure 1. Fringing pattern in VIMOS MOS red data. using real VIMOS MOS data, originally VIMOS fringing problems A small portion of a flat-field exposure obtained with obtained in jitter mode with long slits the LR_red grism is shown, including spectra from six different MOS slits. The red end of the spectra is as part of the VVDS and zCOSMOS sur- The most important characteristic of the towards the top. veys, to simulate a number of different

The Messenger 135 – March 2009 13 Telescopes and Instrumentation Scodeggio M. et al., Improving the Multiplexing of VIMOS MOS

observing strategies, and to evaluate their of the fringing pattern on the CCD pixels Technically, we normalise the flat-field capability of providing a reliable sky sub- which, because of flexure inside the instru- row counts to have the same median traction and an accurate fringing pattern ment, can shift by a few pixels between counts as in the science exposure row removal. The main indication of this work any two exposures. We have seen that it data, and then we compute a very simple is that a data quality comparable to that is possible to compensate for these c2 statistic adding together, pixel by pixel, of VVDS and zCOSMOS data can be offsets by allowing a search for the best- the squared values of the difference obtained with observations carried out in matching fringing pattern between the between normalised flat-field and science stare mode, with relatively small slits. The flat-field and the scientific exposure over exposure. The flat-field row that mini- necessary fringing corrections for the a range of a few pixels (comparable to mises the total c2 value is considered the red spectra can be derived from a flat-field the known extent of the image shift best-matching one. Once such a row exposure, provided that such an expo- resulting from the flexure). This technique has been identified, we scale the total sure is obtained as part of the night-time is schematically described in Figure 2: counts in this row to match those meas- calibrations, immediately before or after for each row of pixels in the image of a ured in the science exposure, and we the wavelength calibration lamp observa- two-dimensional spectrum produced by subtract the scaled flat-field row values tion that is normally executed at the end a MOS slit during a science exposure, from the science exposure row values, of each set of exposures in an Observing we search in the corresponding image effectively producing both a sky subtrac- Block. produced by a flat-field exposure for the tion and a fringing pattern removal in a row that best reproduces the count single step. Any region of the CCD where This flat-field exposure is affected by variations created by the fringing pattern, a significant contribution from the spec- fringing, much like the scientific expo- extending this search over a few rows trum of an object is present on top of sures, except for the precise positioning (typically five) of the flat-field image. the general sky background in the science

1700 Figure 2. Schematic representation of the sky sub- traction and fringing pattern removal technique. 1600 In the centre we show the two-dimensional spectrum 1500 for one MOS slit in a flat-field exposure (left) and in 484 a science exposure (right), obtained with the LR_red 1400 grism. To the right a cut through the science expo- 1300 sure image, at a given row on the CCD (row 482 in this example) is shown. To the left five cuts through 1200 020406080 the flat-field exposure image, for five adjacent rows 1700 around the position of the cut in the science expo- sure are shown. The flat-field row providing the best 1600 fitting pattern to the science exposure data is identi- 1500 fied in red (row 484 in this example). 483 1400

1300

1200 020406080

1700 1700

1600 1600

1500 Line 482 1500 482 1400 1400

1300 1300

1200 1200 020406080 020406080

1700

1600

1500 481 1400

1300

1200 020406080

1700

1600

1500 480 1400

1300

1200 020406080

14 The Messenger 135 – March 2009 Figure 3. A comparison of results obtained using the traditional jitter sequence observation and the stand- ard VIMOS data reduction pipeline v. the new stare mode observation and fringing removal technique. The same two-dimensional sky-subtracted spectrum of a galaxy observed with the LR_red grism as part of the VVDS is shown: the result of the traditional data reduction method (upper); the result of the simulated stare mode observation coupled with the newly pro- posed reduction technique (lower). In this example the new reduction method produces better removal of the fringing pattern than the traditional one. exposure image is excluded from the observation with VIMOS using short- the residual background level (after sky matching pattern search and from the slit masks and stare mode exposures. subtraction) of the two-dimensional spec- rescaling computation. This was approved as programme tra between the red spectral region 281.A-5044 (A), and executed on 26 Sep- (affected by fringing) and the blue region An example of a sky- and fringing- tember 2008. Figures 4 and 5 show (free of fringing). For the data shown in subtracted spectrum is shown in Figure 3, the results obtained by reducing these Figure 4, which effectively still contain the where we compare the results of the data with both the traditional data reduc- whole fringing pattern, the rms variations traditional jitter sequence data reduction tion technique (Figure 4) and the newly in the red part of the background are with the results of the new technique proposed method (Figure 5). It is quite six to eight times as large as those in the proposed here, applied to a simulated clear from a comparison of the two blue part. For the data shown in Figure 5, short slit stare mode observation derived images that the new fringing subtraction reduced with the new fringing removal from a VVDS observation obtained with technique is quite efficient. In order to technique, the rms variations in the red the VIMOS LR_red grism. quantify the impact of the technique bet- part of the background are only two to ter, we have systematically compared four times as large as those in the blue. To the root mean square (rms) variations in visualise the impact that these residuals Validation of new method

Figure 4. Two-dimensional The simulation results briefly discussed sky-subtracted spectra for above indicate that it should be possible a number of slits, produced to obtain good quality spectra for faint using a mask with short astronomical targets from VIMOS spectro­ slits and stare mode obser- vations as part of our DDT scopic observations in MOS mode programme. The blue carried out using a stare mode observing end of the spectra is to the technique. The main advantage of using left. The data have been stare mode observations would be the reduced with the traditional method, appropriate for possibility of using MOS slits that have, on jitter mode observations. average, half the length of the slits used The fringing pattern residu- in traditional jitter mode observations. als are clearly visible in the Using short slits in this way would imme- right half of the image. diately translate into an approximate doubling either of the efficiency or of the total yield for a spectroscopic survey carried out with VIMOS, in comparison Figure 5. Same data as in with the efficiency or yield obtained so far Figure 4, but this time with the VVDS and zCOSMOS surveys. the newly proposed data This result would of course be achievable reduction method has been only for a deep survey, where the number used. The image colour cuts are identical to those in of potential targets is at least twice the previous figure. The as large as the number of slits that could significant removal of the potentially be placed on a MOS mask. fringing pattern, although Given the very significant impact that this not perfect, can be easily appreciated by comparing result could have on future spectroscopic the two images. surveys, we decided to propose a direct on-sky verification of our results.

We therefore submitted a small Director’s Discretionary Time (DDT) proposal to ESO to perform one spectroscopic survey-like

The Messenger 135 – March 2009 15 Telescopes and Instrumentation Scodeggio M. et al., Improving the Multiplexing of VIMOS MOS

Figure 6. Comparison of one-dimensional spectra for one galaxy observed during our DDT run. The top spectrum is extracted from the data shown in Figure 4, and clearly shows the noise introduced by the strong fringing residuals. The bottom spectrum is extracted from the data shown in Figure 5, and shows a more uniform noise pattern in the blue and red half of the wavelength range. s) unit can have on the final one-dimensional ry spectra extracted from these data, we show in Figure 6 a comparison of the extracted one-dimensional spectra for the same object, obtained by starting from the data shown in Figures 4 and 5. Flux (Arbitra It is quite clear from this figure that the noise in the red part of the final spec- tra is significantly reduced by our new data reduction technique. The new data reduction technique is implemented as part of the VIMOS Interactive Pipeline Graphical Interface (VIPGI) data reduction 6 000 7 000 8 000 9 000 pipeline1 (Scodeggio et al., 2005), Wavelength (Angstrom) and will be available to the whole astro- nomical community with VIPGI public extragalactic spectroscopic surveys: will receive a further boost by the planned release 1.4. both the VVDS and zCOSMOS, in their substitution of the VIMOS CCDs, which Wide and Deep samples, have a density should produce spectra with much of potential targets which is two to three reduced fringing residuals in the red part Future prospects times as large as the potential maximum of the spectrum with respect to the number of short slits one could accom- currently installed CCDs. Using the data The results discussed here show that it is modate on VIMOS masks. produced by these new CCDs as input, possible to obtain spectroscopic data our new reduction technique is expected with a quality comparable to that of VVDS The first project to take advantage of this to produce final one-dimensional spectra and zCOSMOS data, which in turn new observing strategy and data reduc- that have uniform noise levels over the translates into a success rate for redshift tion method is already under way: the whole wavelength range covered by the measurement of approximately 90% VIPERS survey (ESO Large Programme VIMOS grisms. This result, coupled for the targeted galaxies, using MOS 182.A-0886 with PI L. Guzzo). From the with the higher quantum efficiency in the masks with short slits coupled with stare set of MOS masks that were designed red part of the spectrum provided by mode observations, irrespective of the for the first semester of observations, we the new CCDs, is expected to increase wavelength range covered by the spectra. can estimate that this survey should be the overall efficiency of a VIMOS-based Using MOS slits that are on average able to observe spectroscopically 60% redshift survey for faint astronomical tar- only half as long as those used for jittered of the parent galaxy sample defined from gets by a factor of two to three. observations would translate into the the complete photometric catalogue. option to place twice as many slits on any In contrast, with a mask design typical of VIMOS MOS mask. To take full advan- jittered observations, and with the same References tage of this option it is necessary to have allocation of telescope time, it would have Le Fèvre, O. et al. 2002, The Messenger, 109, 21 a relatively densely populated starting been possible to sample only some 32% Le Fèvre, O. et al. 2005, The Messenger, 119, 30 sample for the spectroscopic observa- of the galaxies. This increase in efficiency Lilly, S. 2008, The Messenger, 134, 35 tions: our experience with VVDS and has a very significant impact not only Scodeggio, M. et al. 2005, PASP, 117, 1284 zCOSMOS indicates that MOS masks are on the final survey sample, but also on populated with their maximum slit ca­­ the science results expected from this Notes pacity only when the starting sample of project. potential targets includes at least twice 1 http://cosmos.iasf-milano.inaf.it/pandora/vipgi.html as many objects per VIMOS quadrant Finally, the general capability of the ESO as this maximum number of slits. This is community to carry out very large spec- generally the case in relatively deep troscopic surveys in an efficient manner

16 The Messenger 135 – March 2009 Telescopes and Instrumentation

Report on the ESO Workshop Six Years of FLAMES Operations held at ESO Headquarters, Garching, Germany 1–3 December 2008

Claudio Melo1 gurated with UVES almost six years ago, allows the actual mass and radius of Francesca Primas1 the FLAMES community was invited to extrasolar planets to be derived. Transit Luca Pasquini1 participate in an informal workshop held candidates yielded by surveys like OGLE Ferdinando Patat1 at ESO Headquarters from 1–3 Decem- or CoRoT are full of impostors (i.e., dif­ Jonathan Smoker1 ber, 2008. ferent configurations can produce a light curve similar to the one observed in It was a great pleasure to see that almost a genuine planetary transit). Dominique 1 ESO all of the teams who had so far made Naef and Francois Bouchy showed how use of FLAMES attended the workshop. FLAMES contributed to cleaning up The participants were asked to present the list of impostors for transit candidates A significant fraction of the community their scientific results and to add one and to deriving radial velocity curves of users of the VLT multi-fibre spectro­ slide describing the pros and cons of for the very faint OGLE transit planets. graph facility, FLAMES, gathered at ESO using FLAMES to carry out their science. Richard Jackson described a search for Headquarters in December 2008 to These points were collected and used planets and/or binary companions in very present scientific highlights, after in the final open discussion (see below). low mass stars and brown dwarfs in six years of FLAMES operations. This In addition, a few more technical talks which he concluded that the binarity frac- proved to be a great opportunity to were presented by members of the com- tion appears to be lower in these objects review the scientific impact that FLAMES munity and the FLAMES Instrument and than for higher mass stars. has had on different fields of astro­ Oper­ations Team (IOT) presented some physical research and for ESO to assess statistics concerning the use of FLAMES. Age is one of the fundamental parame- the current and future needs of FLAMES During these six years of operations, ters in astrophysics. The Lithium Deple- users. We report on the two and a half about 9000 science frames have been tion Boundary (LDB), i.e., the position day meeting, during which all partici­ taken (with an average of 100 objects in the colour-magnitude diagram that pants openly discussed their experience per image!), the equivalent of 400 nights separates low mass stars with and with- with FLAMES and shared their expertise. of VLT time in total. This corresponds out detected lithium, is a function of to about 25% of the time available on age and is thought to be almost model- Kueyen and is close to the fraction of UT2 independent. Rob Jeffries showed an The Fibre Large Array Multi Element time requested at proposal submission example of how FLAMES can be used to Spectrograph — or simply FLAMES (roughly one third per UT2 instrument). determine the LDB of the faint members — recently completed six years of suc­ of NGC 2547 based on the observations cessful operations attached to the of the Li 670.8-nm line. His approach Very Large Telescope (VLT) Nasmyth A Science highlights could be extended to a handful of clus- focus of Kueyen, at the La Silla Paranal ters that would then constitute bench- Observatory. All participants were invited to give a marks for gauging theoretical isochrones. talk in one of the five different sessions, The combination of an extended field namely, star formation and planets, The study of the Milky Way (MW) Bulge of view with many and varied fibres (a total chemical evolution of the Milky Way and was reviewed by Alvio Renzini, who length of 1.6 km if stretched from end streams, external galaxies, kinematics presented evidence for the very rapid for- to end!) and set-ups has made FLAMES and dark matter, and stellar evolution. mation of the Bulge. Moreover, Bulge a unique facility in ground-based astron- Each session began with an introductory stars seem to be chemically different omy. Thanks to its versatility, FLAMES review talk. All presentations are available from the stars in dwarf spheroidal galax- can be used in many different astronomi- online1. Many of the talks described ies. Although the origin of the Bulge cal applications. Extrasolar planet- samples of stars in the hundreds to thou- has not been fully understood, Renzini hunting, chemical abundances of stellar sands, emphasising the huge multiplex qualified the FLAMES contribution as a groups (globular, open clusters, Galactic gain obtained in using FLAMES over “quantum jump” in this direction. Simone streams, Local Group galaxies, etc.), single-slit instruments. Zaggia presented results of FLAMES kinematics and dark matter, planetary observations in the direction of the Chan- nebulae, the interstellar medium and stel- A few subjectively selected highlights dra Deep Field South where they found lar evolution are only some examples demonstrate the range of scientific ideas more stars than expected beyond 20 kpc of the science cases that have been tar- and prospects presented during the and even one star at 165 kpc. Vanessa geted with FLAMES. workshop; the high quality of the science Hill reviewed the state of our knowledge presentations is of course not limited to of chemical evolution in dwarf spheroidal We thought it was time to celebrate and those cited in this article. (dSph) galaxies and the LMC. Each of to review the performance of FLAMES the LMC, Sagittarius, Fornax, Sculptor during these six years of operations. Fol- The combination of a photometric transit and Carina galaxies shows a distinct lowing the successful experience inau­ with the measurement of radial velocity chemical evolu­tionary track. There is

The Messenger 135 – March 2009 17 Telescopes and Instrumentation Melo C. et al., Report on the ESO Workshop Six Years of FLAMES Operations

some evidence that the abundance pat- wave background and Big Bang nucleo- tigatory study is ongoing. If validated, this tern in metal-poor stars is indistinguisha- synthesis predictions. type of observation could help to improve ble everywhere, although dSph galaxies mass and radius determination of extra- appear to lack the most metal-poor stars Finally, on the extragalatic front, Francois solar planets. It may also provide spin- (Fe/H < –3) found in the MW halo. Hammer reviewed the results on the orbit inclination and indications of addi- morphological and kinematical study of tional (unseen) low-mass companions by Chris Evans and Christophe Martayan galaxies at z ~ 0.6 using the deployable studying variations of the transit time. presented observations of early-type stars Integral Field Units (IFUs). The preliminary in the Magellanic Clouds. In the large conclusions suggest that spiral galaxies Closing the technical session, Francoise programme (LP) described by Evans it are more frequent by a factor of two at Roques proposed a future upgrade of was found that rotational mixing is not as the present day than at z ~ 0.6 (70% ver- FLAMES, aimed at using its large field of dominant as previously thought, with sus 33%). In contrast, the rate of peculiar, view to carry out fast photometry of a both Martayan and Evans finding that low compact or mergers drops from 44% large area of the sky. A dedicated work- metallicity stars spin faster. However, at z ~ 0.6 to about 3% at z = 0, whereas shop on ESO Spectroscopic Surveys2 will a remaining open question is whether the the fraction of luminous infrared galaxies take place in March 2009 at ESO Head- birthplace of a star (in a bound cluster drops from 20% at z ~ 0.6 to 0.5% at quarters in Garching to discuss the future for example) is as important as its initial z = 0. This work is ongoing, with the final of survey instruments such as FLAMES. metallicity in determining the rotational aim being to understand the origin of the parameters. Jonathan Smoker used present-day spirals. archive data from this LP to investigate Open discussion the small-scale structure of high velocity clouds towards the Magellanic system Technical talks The open discussion was moderated by and found variations in Ca ii equivalent Luca Pasquini, who made the initial width of a factor of 10 over a few arc- The morning of the last day was filled with point that, in such a complex instrument, minutes. In another study concerning the more technically-driven talks on data with many different modes and used in gas, Yiannis Tsamis and Alena Zwansig reduction, analysis tools, possible (new) many different applications, the users are de­­scribed the use of ARGUS to map applications and upgrades for FLAMES. quite often the experts. Therefore, the planetary nebulae and protoplanetary exchange of experience, tools, etc. with discs in emission lines at high spatial and In his talk on IFU data reduction, Christer ESO staff and within the community itself spectral resolution to determine the phys- Sandin presented his open software, is desirable, and indeed necessary, to ical properties and chemical abundances used to reduce IFU and ARGUS data, and improve the data quality provided to the of these objects. drew attention to the need to properly users by ESO. account for differential atmospheric Katrin Jordi and colleagues used FLAMES refraction correctly when analysing IFU During the workshop, it became clear to look into the velocity dispersion images. Giuseppina Battaglia described that most FLAMES users do not reduce of Palomar 14. They concluded that their the sky subtraction method developed their data with the ESO pipeline software. results tend to favour more classical by Mike Irwin and applied by her team in Rather, a large fraction still uses the Newtonian mechanics rather than the the chemical study of the dSph galaxies. Geneva Baseline Data Reduction Soft- MOND predictions. In an invited review, ware (BLDRS) that was the only data Gerry Gilmore described the use of The MATISSE package has been devel- reduction software available at the very FLAMES to observe the dynamics of oped by the Nice Observatory team beginning of FLAMES operations. During dwarf spheroidals (supporting flat inner- to analyse stellar spectra to be collected the open discussion, ESO representa- mass profiles) and how FLAMES has by the GAIA mission. A demonstration tives reported that the ESO GIRAFFE been used to resolve the spatial scales of of the enormous potential of the use of pipeline software is now mature, robust, the first enrichment and reionisation. MATISSE to treat FLAMES data was given and produces science-ready data, espe- Thousands of stars have already been by Alejandra Recio-Blanco. Fredric Royer cially when used interactively. Users are observed, although many targets still exist showed a web-based tool designed therefore encouraged to download the for study. to query the FLAMES GTO science-ready package3 and to use it for their own data data obtained by the Observatoire de reduction. Workshop participants men- Andreas Korn put forward evidence for Paris. tioned that the Geneva BLDRS offers the need to include atmospheric diffusion some automated tools to extract further to correctly explain lithium depletion in Luca Pasquini showed the potential of a information from the GIRAFFE spectra metal-poor stars. Indeed, if atmospheric very interesting FLAMES IFU application (e.g., radial velocities, for all science and diffusion is taken into account, the to observe simultaneously the photom­ simultaneous calibration fibres). Repre- observed Li abundances in NGC 6397 etric and spectroscopic transit of a giant sentatives of the ESO FLAMES IOT took can be reconciled with the cosmic micro- planet in front of its host star. An inves­ note of these remarks and will investigate

18 The Messenger 135 – March 2009 solutions, for example improvement of – the accuracy of sky subtraction should Acknowledgment relevant algorithms or adaption of the be quantified and reported in the User We would like to thank all participants for their will- products of the ESO pipeline to interface Manual; ingness and good spirit in sharing results and expe- to relevant data analysis packages. riences during the three days of the workshop. For – the need for a new set of solar spectra us, the ESO staff present, it was a wonderful experi- A number of other technical issues were taken with the new CCD at all settings. ence to share the room with a large fraction of the community of FLAMES users. Finally, we would like brought to the attention of the ESO staff to thank the Director General Discretionary Fund attending the Workshop, including: These and other points that were raised programme for funding this informal workshop. during the Workshop will be discussed – differences in equivalent widths between within the FLAMES IOT for further follow- Notes FLAMES/UVES and FLAMES/GIRAFFE up. The results of these investigations will spectra of the same objects; be disseminated by means of the FLAMES 1 http://www.eso.org/sci/facilities/paranal/instru- webpages4 and related documents. ments/flames/doc/FLAMES_6th_Anniversary/ – typical shifts in the cross-dispersion FLAMES_6th_Anniversary.html 2 http://www.eso.org/sci/meetings/ssw2009/index. direction between the science and html morning flats should be quantified and 3 http://www.eso.org/pipelines reported in the User Manual; 4 http://www.eso.org/sci/facilities/paranal/instru- ments/flames/news.html

The FLAMES facility mounted at the Nasmyth A platform of VLT UT2, Kueyen.

The Messenger 135 – March 2009 19 Astronomical Science

Colour composite image of a region of the Carina Nebula (NGC 3372) formed from exposures in six filters (U, B, V, Rc, Hα and [S ii]) taken with the ESO/MPG 2.2 m telescope on La Silla and the Wide Field Imager. See 05/09 Photo Release for more details.

20 The Messenger 135 – March 2009 Astronomical Science

Studying the Magnetic Properties of Upper Main-sequence Stars with FORS1

Swetlana Hubrig1 2 twin NARVAL at the 2-metre Telescope New magnetic chemically peculiar stars Markus Schöller1 Bernard Lyot on Pic du Midi. ESPaDOnS Maryline Briquet3 and NARVAL can obtain linear and The magnetic chemically peculiar stars Peter De Cat4 cir­cular polarisation spectra at a resolu- with spectral classes A and B (Ap and Thierry Morel5 tion of about 65000. In the southern Bp stars) are presently the best-studied Donald Kurtz6 hemisphere, the visual and near UV FOcal stars in terms of magnetic field strength Vladimir Elkin6 Reducer and low dispersion Spectro­ and magnetic field geometry. Contrary Beate Stelzer7 graph, FORS1, at UT2/Kueyen of the VLT to the case of solar-like stars, their mag- Roald Schnerr8 offers a spectropolarimetric mode with netic fields are dominated by large spatial Carol Grady9 a resolution of up to 4000. The two scales and remain unchanged on yearly Mikhail Pogodin10 11 smaller telescopes, in particular the Tele- timescales. Braithwaite & Spruit (2004) Oliver Schütz1 scope Bernard Lyot, dedicate significant confirmed, through simulations using 3D Michel Curé12 amounts of observing time to magnetic numerical hydrodynamics, the existence Ruslan Yudin10 11 studies by the French and Canadian com- of stable magneto-hydrodynamic config- Gautier Mathys1 munities, making long-term magnetic urations that might account for long-lived, monitoring and magnetic surveys of cer- ordered magnetic fields in these types of tain types of stars possible. However only stars. 1 ESO a few programmes have been devoted 2 Astrophysikalisches Institut Potsdam, to the study of stellar magnetic fields with During 2002–2004, our first survey of a Germany FORS1 in recent years, on account of sample of more than 150 Ap and Bp 3 Instituut voor Sterrenkunde, Katholieke the high demand for observing time with stars, including rapidly oscillating Ap stars Universiteit Leuven, Belgium all instruments installed on Kueyen. (Hubrig et al., 2006a) confirmed that low 4 Koninklijke Sterrenwacht van België, resolution spectropolarimetry of hydrogen Brussel, Belgium One of the biggest advantages of using Balmer lines obtained with FORS1 rep­ 5 Institut d’Astrophysique et de Géophy- FORS1 at an 8-metre telescope is resents a powerful diagnostic tool for the sique, Université de Liège, Belgium the large collecting area, giving high S/N detection of stellar magnetic fields. We 6 Centre for Astrophysics, University of polarimetric spectra of relatively faint first discovered magnetic fields in 63 Ap Central Lancashire, Preston, UK stars, down to magnitudes 12–13. Further, and Bp stars in this survey. Some of 7 INAF-Osservatorio Astronomico di due to the use of Balmer series lines these stars were used for a re-discussion Palermo, Italy for the measurements of stellar magnetic of the evolutionary state of upper main- 8 Institute for Solar Physics, Royal fields in the blue spectral region, ob­­ sequence magnetic stars with accurate Swedish Academy of Sciences, served with grisms 600B or 1200B, fast Hipparcos parallaxes. These new obser- Stockholm, Sweden rotators with v sin i up to 300 km/s can vations confirmed our previous finding 9 Eureka Scientific, Oakland, USA be studied. The technique for measuring that magnetic stars of mass M < 3 M0 10 Pulkovo Observatory, Saint- stellar magnetic fields with FORS1 are concentrated towards the centre of Petersburg, Russia in polarimetric mode was discussed by the main-sequence band, whereas stars 11 Isaac Newton Institute of Chile, Saint- Bagnulo et al. (2001) eight years ago, with masses M > 3 M0 seem to be con- Petersburg Branch, Russia when the first measurement of the well centrated closer to the zero-age main- 12 Departamento de Física y Astronomía, known strongly magnetic chemically sequence (Hubrig et al., 2000; 2007a). Universidad de Valparaíso, Chile peculiar A-type star HD 94660 was dis- cussed in detail. In the course of this study we discovered an extreme magnetic Ap star, HD 154708 We summarise the results of our recent The measurement of magnetic fields (= CD –57°6753), which has the strong- magnetic field studies in upper main- makes use of the presence of circular est longitudinal magnetic field ever sequence stars, which have exploited polarisation in spectral lines, allowing lon- detected in a rapidly oscillating Ap (roAp) the spectropolarimetric capability of gitudinal magnetic field, which is the star, with a mean magnetic field modulus FORS1 at the VLT extensively. component of the magnetic field along < B > = 24.5 ± 1.0 kG (Hubrig et al., the line of sight, averaged over the 2005). This magnetic field is about a fac- stellar disc, to be determined. Since no tor of three stronger than that of Introduction ESO pipeline for the FORS1 spectro­ HD 166473, < B > ≈ 5.5–9.0 kG, the roAp polarimetric mode exists, the spectrum star with the second strongest magnetic Currently, most stellar magnetic field extraction is performed using a pipeline field. In Figure 1 we present recent FORS1 observations are carried out using three written by Thomas Szeifert. The software measurements of the former star over spectropolarimeters. In the northern for measuring the magnetic field strength three months in 2008 used to determine hemisphere, the high resolution spectro­ was developed by us. In the following the rotation period of Prot = 5.367 ± 0.020 polarimeter ESPaDOnS is installed at we describe our magnetic field discover- days. HD 154708 is the first star observed the 3.6-metre Canada France Hawaii Tel- ies achieved with FORS1 in recent years. with FORS1 with a sufficiently uniform escope (CFHT) on Mauna Kea and its phase coverage to establish its magnetic

The Messenger 135 – March 2009 21 Astronomical Science Hubrig S. et al., Magnetic Properties of Upper Main-sequence Stars

8.5 Figure 1. Phase diagram for gap between 0.60 and 0.88, but with the magnetic field measure- somewhat dif­ferent values of the mag- ments of the strongly magnetic star HD 154708; using hydro- netic field strength compared to previous 8 gen and metal lines the measurements obtained with Musicos, best frequency is 0.1863 d-1, ESPaDOnS and NARVAL (Hubrig et al., P = 5.367 days. rot 2008). Unlike FORS1 measurements, 7 . 5 the high resolution spectropolarimeters G) usually do not employ measurements > (k z on hydrogen lines and thus different val-

(< Bz > = 6.326 ± 0.059 kG) was obtained rotation. He i, He ii, C iii, C iv, N ii, N iii and O iii. As an in weather conditions classified as “thick important step, before the assessment clouds”, where the guide star was fre- For a couple of years, the O6Vpe star of the longitudinal magnetic field, we quently lost and thus was repeated by θ1 Ori C remained the only massive exclude all spectral features not belong- the service observer a couple of nights O-type star with a detected magnetic field ing to the stellar photospheres of the later. (Donati et al., 2002). To examine the studied stars: telluric and interstellar fea- potential of FORS1 for the measurements tures, CCD defects, emission lines and of magnetic fields in massive stars, lines with strong P Cygni profiles. This was O stars, pulsating B-type stars and early- in 2007 we obtained 12 observations of the first time that magnetic field strengths type emission line stars θ1 Ori C distributed over the rotational were determined for such a large sample period of 15.4 days and compared them of O-type stars, with an accuracy of a Massive stars usually end their evolution with previous measurements obtained few tens of Gauss, comparable to the with a final supernova explosion, pro­ with high resolution spectropolarimeters. errors obtained with high resolution spec- ducing neutron stars or black holes. The The FORS1 measurements were tropolarimeters. No magnetic fields initial masses of these stars range from suf­ficiently accurate to show a smooth stronger than 300 G were detected in the

9–100 M0 or more, which correspond sinusoidal curve in spite of a phase studied sample, suggesting that large- to spectral types earlier than about B2. The presence of magnetic fields in massive stars has been suspected for a 1.2 1.0 x long time. The discovery of these mag- x netic fields would explain a wide range of 0.9 1.0 Hδ Hγ

0.8 d Flu well-documented enigmatic phenomena, ed Flu ze 0.8 in particular cyclical wind variability, 0.7 H emission variations, chemical peculi- 0.6 α ormali 0.6 arity, narrow X-ray emission lines and Normaliz 0.5 non-thermal radio/X-ray emission. Direct 0.2 0.4 ) meas­urements of the magnetic field )N0.10 0.1 (% (% I strength in massive stars using spectro­ I 0.05 V/ 0.0 V/ 0.00 es es –0.05 ok Figure 2. Left: Stokes I and V spectra of the β –0.1 ok St St Cephei star ξ1 CMa in the blue spectral region –0.10 around high number Balmer lines. Right: Stokes I –0.2 and V spectra of the Be star ο Aqr in the region 3800 3850 3900 3950 4100 4150 4200 4250 43004350 including Hδ and Hγ lines. Wavelength (Å) Wavelength (Å)

22 The Messenger 135 – March 2009 scale, dipole-like magnetic fields with of the 34 SPB stars stud­­ied have been field in the evolutionary process of mass polar magnetic field strengths higher than found to be weakly mag­netic with field ex­­change in a binary system. 1 kG are not widespread among O-type strengths of the order of 100–200 G. stars. Our studies of massive stars revealed In classical Be stars, a number of physi- that the presence of a magnetic field can Two other groups of massive stars, early cal processes (e.g., angular momentum be expected in stars of different classifi- B-type pulsating stars, such as β Cephei transfer to a circumstellar disc, chan- cation categories and at different evolu- and slowly pulsating B (SPB) stars, neling stellar wind matter, accumulation tionary stages. Since magnetic fields can and Be stars, that are defined as rapidly of material in an equatorial disc, etc.) are potentially have a strong impact on the rotating main-sequence stars showing more easily explained if magnetic fields physics and evolution of these stars, it normal B-type spectra with superposed are invoked (e.g., Brown et al., 2004). is critical to answer the principal question Balmer emission, had been a puzzle The magnetic fields of Be stars appear to of the possible origin of such magnetic for a long time with respect to the pres- be very weak, generally of the order of fields. One important step towards ence of magnetic fields in their atmos- 100 G and less. Furthermore, our time- answering this question would be to con- pheres. Only a very few stars of this type resolved magnetic field measurements of duct observations of members of open with very weak magnetic fields had been a few classical Be stars indicate that clusters and associations of different detected before we started our surveys some of them may display a magnetic ages. To date, we have studied the pres- of magnetic fields in B-type stars in 2004. cyclic variability on timescales of tens of ence of magnetic fields only in members Out of the 13 β Cephei stars studied minutes. In Figure 2 (right) we present of the young open cluster NGC 3766 in to date with FORS1, four stars, δ Cet, Stokes I and V spectra of the typical the Carina spiral arm, known for its high ξ1 CMa, 15 CMa and V1449 Aql, possess Be star ο Aqr in the region including Hδ content of early-B type stars, with very weak magnetic fields of the order of a and Hγ lines with a measured longitudinal surprising results. Along with a strong 1 few hundred Gauss. The star ξ CMa is magnetic field < Bz > = +104 ± 33 G. magnetic field detected in a He-peculiar the record holder with the largest mean star, weak magnetic fields have been longitudinal magnetic field of the order of Another emission line star, υ Sgr, is a detected in a few normal B-type stars 300–400 G (Hubrig et al., 2006b; 2009). magnetic variable star, probably on a few and in a few Be stars (Hubrig et al., in In Figure 2 (left) we present Stokes I and months timescale with a maximum longi- preparation). In Figure 4 we present 1 V spectra of ξ CMa in the blue spectral tudinal magnetic field < Bz > = +38 ± 10 G the observed Stokes I and V profiles of region around the high Balmer series (see Figure 3). The evolutionary status for the He-peculiar member of this cluster lines. Distinct Zeeman features, which are this star is not obvious as it is a single-line and of another cluster member that indicators of the presence of a magnetic spectroscopic binary system currently was classified as a potential Be star by field in this star, are easily detected at the observed in the initial rapid phase of mass Shobbrook (1985) with longitudinal mag- wavelengths corresponding to the posi- exchange between the two components netic fields of < Bz > = +1559 ± 38 G tions of strong spectral lines in the Stokes (Koubský et al., 2006). The star is hydro- and < Bz > = –194 ± 62 G, respectively. I spectrum. Figure 2 demonstrates the gen-poor and the observed spectrum great advantage of using the blue- is extremely line-rich (see Figure 3). Future Obviously, to understand the role of mag- optimised EEV2 CCD of FORS1 for mag- monitoring of its magnetic field over a few netic fields in massive stars, future obser- netic field measurements in the blue months with a high resolution spectro­ vations are urgently needed to determine spectral region to cover all H Balmer lines polarimeter would be of extreme interest the fraction of magnetic massive stars from Hβ to the Balmer jump. Roughly half to understand the role of the magnetic and the distribution of their typical field

MJD 54361.08 1.0

1.0 x

) MJD 54343.10

(% 0.8 I ed flu V/ MJD 54333.02

es 0.5 ok 0.6 St Normaliz MJD 53519.91

0.0 0.4 Figure 3. Left: Observed Stokes V spectra of the emission line star υ Sgr over two years in the vicinity 4460 4470 4480 4490 4500 3800 4000 4200 4400 4600 4800 5000 of Mg ii 4481. Right: Normalised FORS1 Stokes I Wavelength (Å) Wavelength (Å) spectrum of υ Sgr.

The Messenger 135 – March 2009 23 Astronomical Science Hubrig S. et al., Magnetic Properties of Upper Main-sequence Stars

strengths. Further, we note that no physi- with age and X-ray emission as expected reported in our earlier studies (Hubrig cal properties are known that define from the decay of a remnant dynamo et al., 2004, 2007b) these lines are very these particular classes of stars as non- (Hubrig et al., 2009, A&A submitted). Dur- likely formed at the base of the stellar magnetic. It seems to be appropriate ing our two-night observing run in May wind, as well as in the accretion gaseous to admit that the inability to detect mag- 2008 we were able to obtain circular flow and frequently display multi- netic fields in massive stars in previous polarisation data for 23 Herbig Ae/Be component complex structures in both studies could be related to the weakness stars and six debris disc stars. No defi- the Stokes V and the Stokes I spectra. of these fields, which can, in some stars, nite detection was achieved for stars with be as little as only a few tens of Gauss debris discs, whereas for Herbig Observations of the disc properties of (e.g., Bouret et al., 2008). Ae/Be stars 12 magnetic field detections intermediate mass Herbig Ae stars were achieved. One of the Herbig Ae suggest a close parallel to T Tauri stars, stars, HD 101412, showed the largest revealing the same size range of the Herbig Ae/Be stars — resolving an magnetic field strength ever measured in discs, similar optical surface brightness enigma intermediate mass pre-main-sequence and structure. It is quite possible that

stars with < Bz > = –454 ± 42 G, confirm- magnetic fields play a crucial role in con- In our recent studies of Herbig Ae/Be ing the previous FORS1 detection by trolling accretion onto, and winds from, stars we sought to expand the sample of Wade et al. (2007). The Stokes I and V Herbig Ae stars, similar to the magnet- stars with measured magnetic fields to spectra of this star are shown in Figure 5. ospheric accretion observed in T Tauri determine whether magnetic field proper- stars. Using our sample of Herbig Ae ties in these stars are correlated with Strong distinct Zeeman features at the stars with masses of 3 M0 or less, we other observed properties such as mass- position of the Ca ii H and K lines searched for a link with other stellar accretion rate, disc inclination, compan- detected in four other Herbig Ae/Be stars parameters to put preliminary constraints ions, silicates, PAHs, or show a correlation are presented in Figure 6. As we already on the mechanism responsible for x 1.0 x 1.0 d Flu 0.8 d Flu 0.8 ze ze 0.6 0.6 ormali ormali 0.4 0.4 0.4 0.2 )N )N 0.2 (% (% 0.1 I I V/ 0.0 V/ 0.0 es es –0.1 Figure 4. Left: Stokes I and V spectra in the blue ok –0.2 ok spectral region around high number Balmer lines St St –0.2 for an He-peculiar member of the young open clus- –0.4 –0.3 ter NGC 3766. Right: Stokes I and V spectra around 3800 3850 3900 3950 3800 3850 3900 3950 high number Balmer lines for a candidate Be star Wavelength (Å) Wavelength (Å) belonging to the young open cluster NGC 3766.

1.0

x 1.0

0.8 x 0.8 d Flu ze 0.6 d Flu 0.6 ze 0.4

ormali 0.4

0.2 ormali 0.6 0.4 )N 0.4 (% )N I 0.2 (% V/

I 0.2 V/ es 0.0

ok 0.0 es Figure 5. Stokes I and V spectra of the Herbig Ae/Be St –0.2 ok –0.2 star HD 101412 with the largest detected magnetic St field.Left: Zeeman features in H9, H8, Ca ii H and K 3850 3900 3950 4000 43004320434043604380 and Hε profiles. Right: Stokes I and V spectra in the Wavelength (Å) Wavelength (Å) vicinity of the Hγ line.

24 The Messenger 135 – March 2009 5 Figure 6. Stokes V spectra in the vicinity of the Ca ii of their pre-main-sequence life (see the H and K lines of the Herbig Ae/Be stars HD 139614, right of Figure 7). These results are HD 190073p HD 144668, HD 152404 and HD 190073. At the top is in line with the conclusions of Hubrig et 4 the previous observation of HD 190073, obtained in May 2005. The amplitude of the Zeeman features in al. (2000, 2007a) that magnetic fields HD 190073 the Ca ii H and K lines observed in our recent meas- in stars with masses less than 3 M are

) 0 urement for this star has decreased by ~0.5% com-

(% 3 rarely found close to the zero-age main-

I pared to the previous observations.

V/ HD 152404 sequence.

es 2 ok

St HD 144668 Closing remarks 1 HD 139614 In summary, using FORS1 spectro­ 0 polarimetric observations, considerable progress has been made over the last 3920 3940 3960 3980 4000 eight years in studies of the presence of Wavelength (Å) magnetic fields in upper main-se­­quence stars. These results are providing sev- eral new clues, but are also posing a number of new open questions requiring 500 500 future spectropolarimetric studies. Currently FORS1 in spectropolarimetric 400 400 mode is the only facility regularly used for observations of circular and linear ) ) (G 300 (G 300 polarisation by the ESO community. Since

>| >| z z FORS1 was decommisioned in P83, the B B polarimetric capability has been moved |< 200 |< 200 to FORS2 with the blue optimised EEV2 CCD available exclusively in visitor mode. 10 0 10 0 We hope that the spectroscopic capa­ bilities of FORS2 will be used in the future 0 0 as intensively and successfully as they 31 30 29 28 27 26 02468101214 were used on FORS1 in the past.

log Lx (erg/s) Age (Myr)

Figure 7. Left: The strength of the longitudinal mag- References netic field plotted against the X-ray luminosity. Right: The strength of the longitudinal magnetic field Bagnulo, S. et al. 2001, The Messenger, 104, 32 as a function of age. Filled circles denote Herbig Ae Bouret, J.-C. et al. 2008, MNRAS, 389, 75 stars with a 3- magnetic field detection, while open σ Braithwaite, J. & Spruit, H. C. 2004, Nature, 431, 819 circles denote Herbig Ae stars with a lower . σ Brown, J. C. et al. 2004, MNRAS, 352, 1061 Squares denote stars with debris discs, of which Donati, J.-F. et al. 2002, MNRAS, 333, 55 none has a 3- magnetic field detection. σ Hubrig, S., North, P. & Mathys, G. 2000, ApJ, 539, 352 Hubrig, S., Schöller, M. & Yudin, R. V. 2004, A&A, magnetospheric activity. For the first decays with age. In Figure 7 (left), we 428, L1 time we established preliminary trends present the strength of the magnetic field Hubrig, S. et al. 2005, A&A, 440, L37 Hubrig, S. et al. 2006a, AN, 327, 289 between magnetic field strength, mass- plotted versus log LX. It is noteworthy that Hubrig, S. et al. 2006b, MNRAS, 369, L61 accretion rate, X-ray emission and age. we find a hint of an increase in the mag- Hubrig, S., North, P. & Schöller, M. 2007a, AN, 328, We find that the range of observed mag- netic field strength with the level of the 475 netic field values is in agreement with the X-ray emission, which suggests that a Hubrig, S. et al. 2007b, A&A, 463, 1039 Hubrig, S. et al. 2008, A&A, 490, 793 expectations from magnetospheric accre- dynamo mechanism may be responsible Hubrig, S. et al. 2009, AN, in press tion models giving support for dipole- for the coronal activity in Herbig Ae stars. Koubský, P. et al. 2006, A&A, 459, 849 like field geometries. Both the magnetic Shobbrook, R. R. 1985, MNRAS, 212, 591 field strength and the X-ray emission There is obviously a trend showing that Wade, G. A. et al. 2007, MNRAS, 376, 1145 show hints of a decline with age in the stronger magnetic fields tend to be found range of 2–14 Myrs probed by our sample, in younger Herbig Ae stars and that mag- supporting a dynamo mechanism that netic fields become very weak at the end

The Messenger 135 – March 2009 25 Astronomical Science

Wolf-Rayet Stars at the Highest Angular Resolution

Florentin Millour1 provide spatially resolved observations leading to these extraordinary events, Olivier Chesneau2 of massive stars and their immediate vicin- although this is not yet firmly established. Thomas Driebe1 ity. However, the number of Wolf-Rayet Only ~ 1% of core-collapse SNe are able Alexis Matter2 stars in the solar vicinity is very low (Van to produce a highly relativistic collimated Werner Schmutz3 der Hucht, 2001; Crowther, 2007). There- outflow and, hence, a GRB. Bruno Lopez2 fore, all Wolf-Rayet stars are too remote Romain G. Petrov2 to be spatially resolved by adaptive optics. WR stars are characterised by an extra­ José H. Groh1 However, for a handful of objects, the ordinary hydrogen-deficient spectrum, Daniel Bonneau2 geometry of the innermost circumstellar dominated by broad emission lines of Luc Dessart4 structures (discs, jets, latitude-dependent highly ionised elements (such as He ii, C iv, Karl-Heinz Hofmann1 winds, or even more complex features) N v or O v i ). The dense stellar winds com- Gerd Weigelt1 can be directly probed with the highest pletely veil the underlying atmosphere spatial resolution available, through the so that an atmospheric analysis can only use of stellar interferometry. be done with dynamical, spherically 1 Max-Planck-Institut für Radioastronomie, extended model atmospheres, such as Bonn, Germany those developed by Schmutz et al. (1989), 2 Observatoire de la Côte d’Azur, Nice, A closer look at Wolf-Rayet stars Hillier & Miller (1998), or Gräfener et al. France (2002). In the two last decades, significant 3 Physikalisches-Meteorologisches Wolf-Rayet (WR) stars begin their life as progress has been achieved in this Observatorium Davos, World Radiation massive objects (usually O-type super- respect, so that WR stars can be placed Center, Switzerland giant stars) with at least 20 times the on the Hertzsprung-Russell Diagram 4 Observatoire de Paris, France mass of the Sun. Their life is brief, and (HRD) with some confidence (see e.g., they die hard, exploding as supernovae Crowther, 2007). and blasting vast amounts of heavy Interferometric observations of high- elements into space, which are recycled Classification of WR stars is based upon mass evolved stars provide new and in later generations of stars and planets. the appearance of optical emission very valuable information of their By the time these massive stars are near lines of different ions of helium, carbon, nature. With the unique capabilities of the end of their short life, during the nitrogen and oxygen. The nitrogen-rich, the VLTI, direct images of their closest characteristic “Wolf-Rayet” phase, they or WN-type, is defined by a spectrum environment where mass loss and develop a fierce stellar wind — a stream in which helium and nitrogen lines (He i–ii dust formation occur, can be obtained. of particles ejected from the stellar sur- and N iii–v) dominate. In the carbon-rich, The breakthrough of the VLTI in terms face by the radiative pressure — expel­- or WC-type, helium, carbon, and oxygen of angular resolution as well as spectral ling mass at a tremendous rate (up to lines (C ii–iv, He i–ii and O iii–v i ) dominate –3 resolution allows competing theoretical 10 M0/yr) while they are synthesising the emission-line spectrum. Castor, models, based on indirect constraints, elements heavier than hydrogen in their Abbott & Klein (1975) demonstrated that to be tested. The high angular resolu- cores. One of the characteristics of WR hot-star winds could be explained by tion made available by the VLTI shows stars is the low hydrogen content of the considering radiation pressure alone. In that there is still a lot to discover about atmosphere due to the stripping-off of this early model, each photon is scattered these massive stars. the outer layers as a result of the strong once at most, and this is sufficient to mass loss. Before and during the WR drive OB-star winds, while more difficulties phase, such stars eject a large amount were encountered for WR stars. Model

Massive stars strongly influence their sur- of matter, more than 10 M0, and with atmosphere studies have advanced suffi- roundings due to their extreme tem­ velocities up to 3000 km/s, which then ciently to enable the determination of perature, luminosity, and mass-loss rate. surrounds the central star in the form of stellar temperatures, luminosities, abun- In addition, they are short-lived, and gas and dust. dances, ionising fluxes and wind proper- their fate is to explode as core-collapse ties of WR stars. What remains uncertain supernovae. Among the extensive zool- Theoretically, the evolution of a WR star are the kinematics of the wind and the ogy of massive stars, Wolf-Rayet stars ends with the collapse of its core and, wind acceleration law. Furthermore, rota- probably represent the last stage of stellar as a rule, the formation of a black hole or tion is very difficult to measure in WR evolution, just before the explosion as a a neutron star occurs. Energetic Type Ib stars, since photospheric features are supernova. So far, Wolf-Rayet stars have and Ic supernovae (SNe) are thought absent. This parameter is crucial when been mainly studied by means of spec- to be direct descendants of massive WR considering GRBs. troscopy and spectropolarimetry, based stars (see, for instance, Crowther, 2007). on spatially unresolved observations. It is now recognised that long-duration Atmospheric models of WR stars are Nonetheless, there is extensive evidence gamma-ray bursts (GRBs) are linked to parameterised by the inner boundary of binarity and geometrical complexity the collapse of massive stars. The merg- radius R∗, at high Rosseland optical of the nearby wind in many Wolf-Rayet ing of the components in a binary sys- depth (typically above 10), but only the stars. In this context, high angular resolu- tem, including a rapidly rotating WR, is optically thin part of the atmosphere tion techniques fill the gap, since they can considered to be one of the channels is seen by the observer. Therefore, the

26 The Messenger 135 – March 2009 determination of R depends on the can be resolved by an interferometer, Probing the wind of the closest WR star: ∗ 2 assumptions made on the velocity law of provided that a minimum spectral resolu- γ Vel the wind, considering that typical WN tion of 500 is available. Therefore, observ- 2 and WC winds have reached a significant ing the LFR of WR stars with long-baseline Among all WR stars, γ Vel is by far the fraction of their terminal velocity before optical interferometry offers the opportu- closest, with a well-known distance today they become optically thin in the continu­ nity to probe their winds in the following of 336 ± 8 pc (North et al., 2007), whilst ­um (especially in the near- and mid-IR). ways: all others are beyond 1 kpc. Owing to its relative proximity, γ2 Vel is relatively bright Typical scales for R∗ are 3–6 R0, depend- ing on the spectral sub-type (Crowther, – By measuring the extension of the LFR and has been studied in great detail, 2007). In the near-IR, the main opacity compared to the continuum. The spatial mainly using spectroscopy. Through 2 comes from free–free interactions, and and kinematical characteristics of the spectro­scopic eyes, γ Vel is shown to be the continuum-forming region is more wind can be better inferred and the wind a binary WR + O system (WC8 + O7.5III, extended at longer wavelengths, reach- velocity law (especially with access to P = 78.53 days), thus offering access to shorter wavelength data) possibly con- the fundamental parameters of the WR ing 2–6 R∗. The core radius of WN stars is larger than of WC stars, but the wind strained. star, usually obtained only indirectly of WC stars is denser, and the continuum through the study of its dense and fast forms farther away from the core radius. – By examining any deviation from wind. The presence of X‑ray emission in As a consequence, the continuum diam- sphericity of these objects. If WR stars the system (Skinner et al., 2001) could be eter is about the same, of the order of were rapid rotators, one would expect explained by an X-ray-emitting bow-shock strong wavelength-dependent devia- between the O-star wind and the WR-star 10–20 R0, corresponding to a diameter of about 0.1–0.2 mas for sources at tions from spherical symmetry, as was wind (the so-called wind–wind collision 2 1 kpc. This implies that the continuum of detected for the luminous blue variable zone, [WWCZ]). γ Vel was observed by both WN- and WC-type WR stars remains star η Car with AMBER/VLTI (Weigelt et the Narrabri intensity interferometer, oper- unresolved for the VLTI in the near-IR, al., 2007) for example. ating at around 0.45 μm, as early as 1968 even with the longest (130 m-scale) base- (Hanbury Brown et al., 1970), but since lines. Yet, this conclusion does not hold – By examining the deviations and per­ the observations lacked spectral resolu- for the line-forming regions (LFR) that can turbations of the WR wind from purely tion, they could not resolve the WWCZ. radial motion. Such deviations may be located at typically 5–50 R∗ and that originate from dust formation very close Our team observed γ2 Vel in December to some of these hot stars. 2004 using the AMBER/VLTI instrument Figure 1. Left: Simulation of the spatial extent of a with the aim of constraining various WC8 star wind in the continuum at 2.10 µm (solid The first, and still most common, method parameters of the system, such as the line) and in three different emission lines (C iii 2.11 µm – dash-dotted line; C iv 2.07 µm – dotted line; C iv of measuring the angular extent of an brightness ratio of the two components, 2.08 µm – dashed line). The apparent size of the WR astrophysical object with a stellar interfer- the spectral and spatial extent of the star in the emission lines can be as much as twice ometer is to use an idealised uniform disc WWCZ, and the angular size associated the continuum size. Right: Simulated images of the to model the stellar photosphere. In real- with both the continuum and the lines same star in the continuum region around 2.10 µm (centre) and in the C iv 2.07 µm line (right). The ity, WR stars are not “hard-edged” in emitted by the WR star. AMBER delivered continuum can be described by a Gaussian intensity the continuum, which forms directly in the spectrally dispersed visibilities as well distribution, while the superposition of continuum wind (see Figure 1). as differential and closure phases and, of and line flux leads to an intensity distribution that can be described by a limb-darkened disc.

p (R�) 0102030 2.5

2.0

1.5 ed flux unit

1.0 Normaliz

0.5

0.0 024681012 p (R ) ∗

The Messenger 135 – March 2009 27 Astronomical Science Millour F. et al., Wolf-Rayet Stars at the Highest Angular Resolution

course, also a spectrum of the object. since we observed a separation of independent spectrum of the WR com- +0 .11 2 During these observations, AMBER 3.62 –0.30 mas, implying a distance of ponent of γ Vel in the K-band. This +38 worked with a resolution of R = 1500 in 368 –13 pc, which is in agreement with allowed us to compare the WR spectrum the spectral band 1.95–2.17 μm. We inter- recent spectrophotometric estimates. with line-blanketed radiative transfer preted the AMBER data in the context An independent observation by the models of WR stars (Dessart et al., 2000). of a binary system with unresolved com- Sydney University Stellar Interferometer The match between the inferred WR ponents, neglecting, to a first approxi­ (SUSI) confirmed and refined the distance spectrum and the modelled spectrum is 2 +8 mation, the wind–wind collision zone flux estimate of γ Vel at 336 −7 pc (North et contri­bution and the extension of the al., 2007). WR wind (Millour et al., 2007). Based on Figure 2. γ2 Vel as seen by AMBER in 2004 (Millour the accurate spectroscopic orbit and In contrast to SUSI, which uses broad- et al., 2007): the observations (lower plots) are fitted the Hipparcos distance (258 +41 pc), the band filters, AMBER allowed us to dis- successfully by a model involving a WR star and an –31 O star (grey lines). One can extract the spectrum of expected separation at the time of the perse the K-band light with a resolution of the Wolf-Rayet star alone from this model-fitting observations was 5.1 ± 0.9 mas. How- 1500. Therefore, we were able to sepa- (top-right panel, blue curve) and compare it with a ever, our observations showed that rate the spectra of the two components; radiatively-driven WR wind model spectrum (red the Hipparcos distance was incorrect i.e., we obtained, for the first time, an curve). The general view of the system (top-left) also involves a wind–wind collision zone.

4 E

N

P ) 1 mas unit

ry 2 WR star ra rbit (a

0 Flambda 7 5 9 9 0 1 9 4 8 03 16 05 11 01 07 07 13 Ω 10 2. 2. 2. 2. 2. 2. 2. 2.

2.

II I I II

I IV IV IV II IV He C He C C C He He C O star Wind-wind collision zone 2.00 2.05 2.10 2.15 Wavelength (µm) Wind blown cavity Background image: DSS2 Orbital Phase 0.32

6 UT2−UT3 UT2−UT3 1 0 ) ) ad ty (r ad 4 (r

UT3−UT4 ase UT3−UT4 Ph

al Visibili 0 –2 ti al ti en en er ff er 2 ff Di UT2−UT4 Closure Phase UT2−UT4 Di −1 –4

0 2.04 2.06 2.08 2.10 2.12 2.04 2.06 2.08 2.10 2.12 2.04 2.06 2.08 2.10 2.12 Wavelength (µm) Wavelength (µm) Wavelength (µm)

28 The Messenger 135 – March 2009 relatively good, except for some regions this additional light source in the system law with a steep inner region and a flat where the match is worse (especially in a between the two main stars. If this outer region. Leaving aside the technical few emission lines and in some parts source is confirmed, it would be located details, this basically implies that different of the continuum, see Figure 2). This sug- closer to the O star than to the WR velocity laws predict a different extent gests that as a second order perturba- star and would contribute approximately of the LFR. Therefore, we are currently tion, an additional light source in the sys- 5% of the total flux of the system (i.e., continuing our investigations on γ2 Vel tem may affect our fit. We checked that additional free–free emission). This is in with the VLTI to disentangle better the this marginal mismatch cannot come from agreement with the expectations from different light contributions in this fasci- a flat continuum emission, and therefore, Millour et al. (2007) and, possibly, we nating system. we favoured the presence of an addi- have detected the wind–wind collision tional compact light source in the system. zone (WWCZ) between the two stars. The dust formation puzzle After that first study, we carried out follow- However, other effects, such as the LFR up observations of γ2 Vel in 2006 and extent, can also affect our interferometric One challenging problem related to WR 2007. In addition to a refinement of the signal. Indeed, in the past, different veloc- stars is understanding how dust can form orbital solution of North et al., (2007), ity laws were used for models of γ2 Vel, in the hostile environment of these hot Millour et al. (2008) were able to locate e.g., a steep velocity law or a combined stars. It turns out that several of the dusty WR stars harbour a so-called pinwheel nebula, as shown recently by the obser- Figure 3. What kind of visibility sig- nature can we expect from a pin- vations of Tuthill et al. (see, for example, 60 wheel nebula? This example using their article in 2008). From an interfero- a pinwheel “toy model” (top-left) metric point of view, pinwheel nebulae 40 illustrates the whirlpool shape of appear in the visibilities as typical whirl- the 2D visibility map (bottom-left) as well as the plateau-like shape pool modulations that can be detected 20 of the visibilities towards longer with the current performance of the VLTI ) baselines, as seen from the radial (see Figure 3). When the number of data- as cuts in the bottom-right panel. A sets is very limited, clues to the presence 0 two-component Gaussian is also (m shown for comparison. These two of a pinwheel can also come from a pla- δ characteristic visibility signatures teau-like shape in the visibilities towards −20 provide clues to the presence of longer baselines. Of course the detection a pinwheel around a dusty WR of new pinwheel nebulae would provide star. −40 a more direct evidence of the binarity of the observed stars, but a clear detection −60 of the underlying binary in known pin- wheels is also one of the goals of study- ing these dusty WR stars. Up to now, we −60−40 −20020 40 60 have successfully observed a few dusty α (mas)

1.0 150

100 0.8

50 0.6 ) ty

(m 0 isibili V V 0.4 −50

−100 0.2 −150

050100 150 −150 −100 −50050 10 0150 Baseline length (m) U (m)

The Messenger 135 – March 2009 29 Astronomical Science Millour F. et al., Wolf-Rayet Stars at the Highest Angular Resolution

1. 0 0.4

0.3 ty ty 0.2 isibili isibili V V 0.5

0.1

0.0 0.0 050100 15 0 050100 Base length (m) Spatial frequency (as–1) 0.15

Figure 4. Examples where pinwheel signatures in WR stars may have been detected by the VLTI. Top-left: WR48a-b, together with the same pinwheel “toy model” as in Figure 3. The observed data are 0.10 marginally matched by a Gaussian shape (full line) and may be better explained by the pinwheel model ty (dashed lines). Top-right: WR118, where a two- Gaussian model (red line) does not match the AMBER

isibili data (black lines) at longer baselines. In particular, V the squared visibility bump and the non-zero closure 0.05 phase (not shown) are strong hints that a pinwheel also resides in this system. Lower panel: The MIDI measurements of WR104 show the typical visibility modulation from the already known pinwheel. The different curves represent different VLTI baseline 0.00 lengths and orientations. 10 15 20 25 30 35 40 45 Spatial frequency (as–1)

Wolf-Rayet stars with the VLTI, which AMBER/VLTI instrument. Both objects remarkable changes in the dust emission illustrates the potential of optical/IR inter- have angular diameters of about 4–6 mas have been observed in the past two dec- ferometry to contribute to the fascinating in the K-band and mid-IR sizes of the ades, WR 118 is classified as a permanent question of the binarity of these stars. order of 10–15 mas. The sizes measured dust producer. The extended dust enve- are quite large, compared to the ex­pected lope of WR 118 was successfully resolved WR 48a is a WC star producing dust size of single WR stars (0.1–0.5 mas), for the first time by Yudin et al. (2001) in the form of eruptions, which suggests and these two objects are probably two using bispectrum speckle interferometry the presence of a companion. It is found good candidates to harbour pinwheel with the 6-metre BTA (Large Altazimuth to be within 1 arcsecond of the two nebulae, even though no clear signature Telescope). Yudin et al. concluded that the heavily reddened, optically visible clusters has been formally detected yet (Figure 4). apparent diameter of WR 118’s inner dust Danks 1 and 2, which are themselves These two targets are definitely inter­ shell boundary is 17 ± 1 mas. We recently separated by only 2 arcseconds. Another esting because they produce dust and, measured WR 118 with AMBER in low- object that also seems to be a dusty hence, are likely binary candidates, spectral resolution mode (see Figure 4). WR star (MSX6C G305.4013 + 00.0170) belonging to a young star-forming region At first glance, the AMBER visibilities sug- was found close to WR48a, which we whose distance is more accurately known gest that there is an unresolved compo- hereafter denote as WR48a-b. No optical than is usual for WR distances. nent contributing approximately 15% to counterpart is reported in the literature the total K-band flux. In addition, one can and the JHK photometry for this source WR 118 is a highly evolved, carbon-rich see that the K‑band visibility is not spheri- from the Two Micron All Sky Survey Wolf-Rayet star of spectral type WC10. cally symmetric, and we detected non- (2MASS) shows a spectral shape indica- It is the third brightest Wolf-Rayet star zero closure phases. These clues illustrate tive of an object with high extinction in the K-band (K = 3.65), and its large IR- that the overall shape of the measured

(AV ~ 10 mag). WR48a and WR 48a-b excess is attributed to an envelope com- visibility of WR 118 is in qualitative agree- were successfully observed with the posed of carbonaceous dust. Since no ment with the visibility signature expected

30 The Messenger 135 – March 2009 from a spiral-like dust distribution. Cur- geometry and the ephemeris of this lines. Observing with shorter wavelengths rently, the interpretation of the recently source were not accurately known. These (H or J bands, or even visible), would obtained AMBER data is being subjected observations promise to evaluate pre- boost this research field, as both the inter- to a more detailed analysis. cisely the size of the dust-forming region ferometer resolution and the LFR extent and the dust production of the system, as increase towards shorter wavelengths. The most recent dusty WR star observed viewed at high spatial resolution. This would open the possibility to directly with the VLTI so far was WR 104 with validate the competing models for WR the MIDI instrument. WR104 is the arche- winds or, maybe, could even lead to new type of the colliding-wind binary creating Prospects and unexpected results. a beautiful pinwheel system, which was first detected by the technique of The studies described correspond aperture masking with the Keck tele- to work in progress with the VLTI for References scope (Tuthill et al., 2008). We observed observing WR stars. No imaging of these Crowther, P. A. 2007, ARA&A, 45, 177 it with MIDI using two of the auxiliary sources has been done up to now, but Castor, J. I. et al. 1975, ApJ, 195, 157 telescopes. The visibilities show the sig- the AMBER instrument, with three tele- Dessart, L. et al. 2000, MNRAS, 315, 407 nature of two flux distributions on the scopes combined, has some imaging Gräfener, G. et al. 2002, A&A, 2002, 387, 244 source: a Gaussian-shaped dusty enve- capabilities. A first image of a pinwheel, Hanbury Brown, R. et al. 1970, MNRAS, 148, 103 Hillier, D. J. & Miller, D. L. 1998, ApJ, 496, 407 lope surrounding the binary system with resolutions much higher than single Millour, F. et al. 2007, A&A, 464, 107 and producing 80–90% of the flux in the pupil telescopes, would be a milestone in Millour, F. et al. 2008, SPIE, 7013 mid-infrared, and the dust pinwheel. IR long-baseline interferometric research North, J. R. et al. 2007, MNRAS, 377, 415 The characteristic pinwheel signal is seen on dusty WR stars and would probably Schmutz, W. et al. 1989, A&A, 210, 236 Skinner S.L. et al. 2001, ApJL, 558, 113 by MIDI as a cosine modulation, evolving trigger more VLTI observing programmes Tuthill, P. G. et al. 2008, ApJ, 675, 698 with time as the pinwheel rotates and in this field. On the other hand, reaching van der Hucht, K. A. 2001, New Astronomy Review, the interferometer baseline length and ori- sufficient dynamic range with AMBER/ 45, 135 entation change (see Figure 4). Such VLTI in medium spectral resolution mode, Weigelt, G. et al. 2007, A&A, 464, 87 Yudin, B. et al. 2001, A&A, 379, 229 a limited amount of data would be ex­­ would allow one to directly constrain the tremely difficult to interpret if the global WR wind velocity field in the emission

Image of Jupiter taken in near-infrared light with the VLT Multi-conjugate Adaptive optics Demonstrator (MAD) prototype instrument on 17 August 2008. This colour composite is formed from a series of images taken over a time span of about 20 minutes, through three filters at 2.2 µm (Ks-band), 2.14 µm (Brackett-γ continuum) and 2.17 µm (Brackett-γ). The filters select hydrogen and methane absorption bands. The spatial resolution is about 90 milliarcsec- onds across the whole disc, corresponding to 300 km on the planet surface. See ESO PR 33/08 for more details

The Messenger 135 – March 2009 31 Astronomical Science

The Beauty of Speed

Andrea Richichi1 the highest spectral resolution or the re­­quests for fast observations are starting Cesare Barbieri2 utmost sensitivity, and these demand in to be implemented on a best-effort Octavi Fors3 4 turn long integrations. basis. Today, several instruments at the Elena Mason1 La Silla Paranal Observatory offer high Giampiero Naletto5 Yet there is a wealth of knowledge to be time resolution, as summarised in Table 2. gained by going to the other extreme, We have not included here subsystems and observing with high time resolution that by necessity have to include fast 1 ESO Garching (see Table 1). Pulsars, stellar pulsations operation, such as adaptive optics and 2 Dipartimento di Astronomia, Università and oscillations, flares and bursts, tran- fringe trackers. di Padova, Italy sits and occultations, and more, are phe- 3 Departament d’Astronomia i Meteoro­ nomena that are best studied by record- Modern panoramic detectors have quite logia, Universitat de Barcelona, Spain ing data at rates much faster than those large formats, and as a consequence 4 Observatori Fabra, Barcelona, Spain usually employed by astronomers. The typical readout times are, at minimum, 5 Dipartimento di Ingegneria extragalactic community need not feel of the order of seconds. In order to beat dell’Informazione, Università di Padova, left out either: for example, the variability this limit by up to three orders of magni- Italy of active galactic nuclei (AGN) holds a tudes, as required by some applications, crucial key to the size and structure of the it is gen­­erally necessary to sacrifice the central engine, and, if it is studied on number of pixels, by reading out only a The burst mode of ISAAC has been used timescales of minutes today, it is natural small sub-window (used for example by systematically to record lunar occulta- to expect that in the era of Extremely the ESO infrared [IR] instruments ISAAC, tions with high time resolution, produc- Large Telescopes (ELTs) these timescales SOFI and NACO). Other approaches are ing several unique new results that could be down to seconds. Last but not to shift the charge in CCDs, as in FORS2. remain unattainable by any other tech- least, if we really observe very fast, then Other detectors are intrinsically quite nique. This is not the only possible we can also beat atmospheric turbulence, fast by design and have a relatively small choice of instrument for high time reso- to the point that, at least for some ap­­ format, such as the mid-IR detectors of lution, and fast time modes of one kind plications, we no longer need expensive VISIR and MIDI that have to avoid satura- or another have been implemented and complex correction systems. tion by the high background signal. on several other ESO instruments. We provide a brief overview of the present To be sure, there are many problems in In the following we will focus on recent capabilities and summarise some sci- going very fast: we collect far fewer pho- results in two areas: the near-IR detection entific results. We speculate about the tons; we have to fight harder against of lunar occultations with ISAAC in the future of high temporal resolution appli- detector noise and other unpleasant fea- millisecond range; and the “blazingly fast” cations, presenting the trail-blazing tures; there may not be sufficient time detection enabled by a unique instrument instrument Iqueye that recently com- to read the whole area of our large-format that recently had its first technical run in pleted its first technical run at La Silla. detectors for which we have paid so La Silla. dearly. But first and foremost, what do we mean by “fast”? Examining Table 1, it The quest for high time resolution will be noticed that the wish-list spans a At Paranal: the Moon in slow motion range of six orders of magnitudes or Encoded somewhere in the human more, from seconds down to microsec- Lunar occultations (LO) are a phenome- ge­­nome, there must be a love for speed. onds or less. At ESO, most of the in­­ non in which the lunar limb acts as a Filippo Marinetti, founder of the artistic struments currently in operation were straight diffracting edge. As the Moon movement known as Futurism, stated it originally designed without an explicit moves over a distant background source, best in his Manifesto of 1909: “We affirm requirement for high time resolution. But a fringe pattern is generated that moves that the world’s magnificence has been as the standard modes have become over the observer. Typical speeds of enriched by a new beauty: the beauty more and more routine, a number of this pattern are about 0.5–1 m/ms (or of speed... Time and Space died yester- day. We already live in the absolute, Phenomenon Timescale (current) Timescale (ELT era) because we have created eternal, omni- Stellar flares and pulsations seconds, minutes 10–100 ms present speed.” Young or old, many peo- Stellar surface oscillations 1–1000 μs 1–1000 μs ple are infatuated with the concept of e.g., white dwarfs, neutron stars 0.1 μs speed: in technology, in sport, or on the Tomography, eclipses, flickering 10–100 ms 1–10 ms e.g., close binary systems highway. But what about astronomers? Pulsars 1 μs–100 ms 1 ms–1 ns? It is gen­erally considered that, in spite of Variability in AGN minutes seconds? some peculiarities, they are people too, Stellar occultations 1 ms 1 ms and therefore they should not be immune Planetary occultations and transits 100 ms–10 s 100 ms–10 s to the fascination with speed. However, almost all the instruments that astrono- Table 1. Fast time applications in astronomy (partly mers design and build are geared to based on an E-ELT study by Redfern & Ryan, 2006).

32 The Messenger 135 – March 2009 Instrument Modes Detector Time Rate (Window) Configuration and Mode VISIR Burst DRS 12.5 ms SF imaging, visitor SOFI Burst, FastPhot Hawaii 4 ms (8 x 8), 15 ms (32 x 32) imaging, visitor ISAAC Burst, FastPhot Hawaii-1, Aladdin 3 ms (32 x 32), 6 ms (64 x 64) imaging, visitor, service ISAAC Burst Hawaii-1, Aladdin 9 ms (1024 x16) spectro, under commisioning NACO Cube Aladdin 7.2 ms (64 x 64), 350 ms (1024 x 1024) imaging, visitor HAWK-I Fast Hawaii-2RG 6.3 ms (16 x 16) imaging FORS2 HIT CCD (charge shift) up to 2.3 ms image/spec, visitor, service VLTI Fast Various up to 1 ms image/spec, not foreseen

500–1000 m/s). The fringe spacing is an angular resolution as good as 0.5 mil- Table 2. A summary of instruments available at ESO determined by the distance to the Moon liarcseconds (mas), or about 100 times for fast time resolution with their main characteristics. and the wavelength of observation: in better than the diffraction limit of the tele- the near-IR it is a few metres, so that time scope and comparable with that of the maser source, we could derive a clear sampling of about one millisecond is much larger and complex VLTI; a limiting detection of the circumstellar shell and an required to measure the fringe pattern. sensitivity close to K ≈ 12.5, or several estimate of its angular size and distance. From this measurement, a wealth of infor- magnitudes fainter than the VLTI; and a As a result of these first observations, mation on the background source can dynamic range of 8 magnitudes even the so-called burst mode of ISAAC has be recovered, including the angular diam- within one Airy disc from the central star. subsequently been offered on a regular eter of stars, the projected separation All this in the blink of an eye, or, to basis since Period 80, and is now a and brightness ratio of binaries. It is also be realistic, in a few minutes, taking into routine technique for LO observations at possible to reconstruct the brightness account telescope pointing and data the VLT. Since LO observations require profiles of complex sources by a model- storage: sounds too good to be true? a minimum amount of time, they represent independent analysis. A typical concern There are indeed some major limitations an ideal filler programme: every interval of of those who learn about LO for the first to LO: they are fixed time events, and at least five minutes during which no other time is the effect of mountains and irregu- we cannot choose the targets at will. service mode programmes are available larities of the lunar limb. Luckily, this can generally be safely neglected, since we We report a summary of the results are dealing with diffraction and not with obtained in those first two runs in Table 3. Table 3. Statistics of the results obtained from the first burst-mode runs with ISAAC for lunar occulta- geometrical optics. A number of peculiar- Figure 1 provides an illustration: for the tions. R and D events are reappearances and disap- ities characterise LO, in particular the AGB star 2MASS 17453224-2833429, a pearances, respectively. fact that the angular resolution achieved is not dependent on the size of the tele- March 06 August 06 Total scope used for the observation. In fact, Total hours 4.2 8.5 12.7 the Moon itself can be considered as our Type of event R D telescope in this case. Attempted events 51 78 129 Successful events 30 72 102 We have already reported in a previous Diameters 3 1 4 Messenger article (Richichi et al., 2006) Binaries/triples 2/0 6/1 9 on the first successful observations of Shells/complex 0 2 2 LO at the Very Large Telescope (VLT) Planetary nebula, central stars 0 1 1 with ISAAC at Unit Telescope 1. Using Masers 2 1 3 the Aladdin detector, an area of 32 x 32 pixels were read out every 3.2 ms. We 250 announced this ground-breaking per- 200

formance at the time, and this has now le been confirmed by the detailed analysis y 150 nsit carried out in two recently published Profi 0.05

te 100 ss papers (Richichi et al., 2008a; 2008b): In 50 htne 0 Figure 1. Left: Upper panel, data (dots) and best fit 20 (solid line) for 2MASS 17453224-2833429. The lower . Brig panel shows, on a scale enlarged by four and dis- x4 10 placed by arbitrary offsets for clarity, the residuals of

0 Norm three different fits. The upper two are for an unre- solved and a resolved stellar disc (reduced χ2 = 6.3), –10 the lower one is with a model-independent analysis 0 Residuals –20 (reduced χ2 = 1.6). Right: Brightness profile recon- structed by the model-independent analysis. The 1100 1200 1300 1400 1500 –20020 inner shell radius is estimated to be about 40 AU. Relative Time (ms) Milliarcseconds

The Messenger 135 – March 2009 33 Astronomical Science Richichi A. et al., The Beauty of Speed

Figure 2. Same as Figure 1, for a source observed in Period 81 (K = 7.1, no known bibliographical entries). The fit residuals are for a point source (upper) and a binary source with a separation of 9.4 mas and a brightness ratio of 1:11 (lower). The right panel shows the reconstruction by a model-independent method. or for which the atmospheric conditions 1250 are not met, an LO observation can be attempted if the Moon is above the hori-

A le zon. In fact, we note that LO observations y 1200 B nsit are almost insensitive to seeing and other Profi 0.05 te

adverse atmospheric conditions. With ss In 1100 this strategy in mind, we submitted a filler proposal for Period 80, which unfortu- htne nately did not produce many results due

10 . Brig to the unavailability of ISAAC during x4 much of the period. However we submit- ted again for Period 81 and we waited 0 Norm during Period 82 so we could see the first 0 results. These were very encouraging, Residuals –10 and the programme was resubmitted for 900 1000 1100 1200 1300 1400 –20020 Period 83, and accepted. We provide Relative Time (ms) Milliarcseconds here a first account of the observations carried out in Period 81. (viz. sky image with time). The masks iso- faint magnitudes, make this a valuable list late the signal from the star, which, at of calibrators for long-baseline interferom- For Period 81, we computed LO predic- these data rates, often has a variable and etry, starting with the VLTI. tions to a limiting magnitude of K = 9.3 aberrated image, using twin criteria of — a compromise between the number of contiguity and continuity. The background computations and the volume of potential is then computed from the remaining At La Silla: zooming in on pulsars events. This choice resulted in 28 682 pixels and subtracted. The pipeline proc- events observable within good observa- esses the resulting light curve further The observations presented so far may tional constraints. Note that we discard using a multi-resolution wavelet transform seem fast by the standards of most full and waning Moon phases, the first analysis to produce first guesses of astronomers, but they pale in comparison because of the brightness and the latter parameters such as the time of occulta- with some observations recently carried because they are more challenging, and tion, the intensity of the star, etc. From out in La Silla. In January 2009, a new not well suited to service mode. We these parameters an initial fit is produced, instrument emerged from its packing developed a prioritisation rule based on and we are then presented with prelimi- cases, and was quickly assembled at the the K magnitude and J–K colour, prefer- nary results and quality estimates that Nasmyth B focus of the New Technology ring stars with very red colours indicative allow us to carry out an interactive analy- Telescope (NTT) by a team of Italian of possible circumstellar extinction. We sis more efficiently and to focus on the astronomers and engineers. Iqueye (Fig- further refined the priorities for sources interesting cases. First indications are that ure 3) is the NTT version of Aqueye, a with known counterparts or that had about six stars observed in Period 81 prototype already previously deployed at been studied previously. Then we applied appear to be binary, with projected sepa- the Asiago Observatory. The “queye” part a selection rule such that for every five rations as small as 5 mas. One of them of their names indicates the close rela- minute interval only one star would be is illustrated in Figure 2. A few stars also tionship to QuantumEye, an instrument selected, with the highest priority among appear to be resolved, although a more concept proposed initially for OWL, and the other events close in time. As a result, detailed analysis is still needed. now for the E-ELT (Dravins et al., 2005; 1629 Observing Blocks (OBs) were gener- Barbieri et al., 2008). The driver for this ated — a significant load for the User Most of these results, as well as those class of instruments is to reach ultimately Support Department and our colleagues obtained previously, pertain to stars that the regime of time resolution in which at Paranal to handle, whom we thank for are absent from the literature or have photons are subject to quantum limits. their support. Equipped with this reservoir little information available. For some of From Figure 4, it can be seen that the of OBs, our programmes lay dormant until them, not even an optical counterpart Heisenberg uncertainty principle plays a summoned by an ISAAC night astrono- is known. Clearly, further investigations dominant role in the visible range when mer… an exciting wait for us in Garching! are needed in order to collect direct the time resolution approaches a few By the end of the period, a total of 125 LO imaging by adaptive optics and optical picoseconds (1 ps = 10–12 s). But how is it events had been attempted in service and near-infrared photometry, and to possible to reach such time resolutions? mode. Of these, 116 resulted in positive provide spectral classification. We are detections of well-recorded light curves, a waiting to prepare a corresponding pro- The solution selected for Iqueye is to com- very satisfactory outcome indeed. posal when a significant number of such bine astronomy with the state of the art stars are available. At the same time, we offered by detector and nuclear physics. In order to cope with such a volume note that a large number of unresolved Iqueye and its siblings are equipped with of data, we developed our own data pipe- sources have been measured with upper single photon counting avalanche photo­ line (Fors et al., 2008), which generates limits on their angular diameters of order diodes (SPADs) which attain 50 ps time extraction masks for each data cube 1 mas. These limits, and the relatively resolution with count rates as high as

34 The Messenger 135 – March 2009 Figure 3. Iqueye at the NTT, with some members of the team and of the supporting staff from La Silla. Two of the four SPADs are clearly seen, protruding Figure 4. Astronomy in time and frequency. Pushing from above and below the middle of the instrument. the time resolution towards the limits imposed The rack includes the control electronics and the by the Heisenberg uncertainty principle might be rubidium clock. compared with the opening of a new window.

Log electromagnetic frequency (Hz) 22 20 18 16 14 12 10 8

GammaX-raysUVVisibleInfraredRadio 20 Photon counting No photon counting 15

10 l (s)

5 Crab Nebula

Classica Domain Giant Pulses

escale 0 T = 10 37 K 1 ms –5 g tim m tu Lo –10 1 ns

1 ps Quan Domain s n physic –15 t = 1/ν know n from tum limit ∆ rmatio –20 Quan l info No additiona

–14–12 –10–8–6–4–20 Log electromagnetic wavelength (m)

10 MHz. SPADs are very robust to high board originally designed for CERN and a The basic scheme of Iqueye is shown in light levels, use standard voltages and do reliable and accurate time base. For this Figure 5 and a more detailed account will not require external cooling; their quantum latter task, both a Global Positioning Sys- be presented in the near future. For the efficiency reaches 60% in the visible. One tem (GPS) signal and a rubidium clock moment, we show some impressive light shortcoming of these detectors is their are employed. This clock has the required curves of pulsars (Figures 6–8). The first deadtime, which at 70 ns limits the count resolution, but it is not sufficiently stable, was obtained with Aqueye at the Asiago rates to 14 MHz. Such devices are in a and so it is aligned occasionally with 1.8-metre telescope in October 2008 constant state of evolution, and improve- the GPS signal. By combining these two (average signal over 30 minutes), and is ments are foreseen in the near future. systems, the time-tagging achieved in compared with that taken at a much larger Both to avoid saturation in case of bright Iqueye is reliable to about 30 ns over long telescope. At the NTT, Iqueye was able sources, and to provide independent periods of time, and to much less over to detect the Crab pulsar in a single cycle, detection statistics, Iqueye splits the tele- shorter periods. The capabilities of Iqueye showing its individual pulses clearly. The scope pupil into four, and each quadrant are impressive, especially if one considers quality is unprecedented, and theoreti- is observed with one SPAD. However that the whole is achieved in a highly com- cians will be able to zoom in on small this is only one part of the problem: the pact and portable instrument, installed details and refine their models. One can other gigantic hurdle is how to tag the and successfully operated at the NTT by a thus study pulse amplitude variations, arrival time of each photon. This has been small group of people, most of whom had noise in the arrival times, precession of achieved with two aids: an event-tagging never visited the site before! the rotation axis, correlations with the giant radio pulses and much more. But the Crab pulsar is not the limit: at the NTT, Centering Global filters Iqueye has been able to study another Camera wheel pulsar, B0540-69, with a V magnitude of Mirror Sliding probe only 23! This pulsar had previously been pin-hole Telescope mirror measured with good quality only by the focus Hubble Space Telescope (HST), and com- parison of the NTT and HST data will e Pyramid provide an accurate deceleration rate for mirror the pulsar. Note that these results are lescop still preliminary, and the quantities in the Te figures need to be corrected for a number Variable Focal of effects. Removable reducer pin-hole

neutral filter e lik - Single arm Figure 5. Schematic representation of Iqueye. Both 4) YE the centering camera and the individual filters in filters wheel (x Field Camera uE each of the four arms are planned, but could not be

single arm included in the version for the first run at La Silla. AQ With these additions the efficiency is increased and can provide the capability for simultaneous multi­ SPAD colour observations.

The Messenger 135 – March 2009 35 Astronomical Science Richichi A. et al., The Beauty of Speed

Figure 6. The light curve also offers a new approach to the tech- 250000 of the Crab pulsar (peri­ nique known as intensity interferometry, ­od 33 ms) observed with Aqueye in 2008 (blue, originally devised and applied by 200000 1.8-metre telescope in Hanbury-Brown and colleagues. The Asiago) and with the Kitt method revolutionised interferometry in

te Peak 4-metre telescope the 1960s through the study of the 150000 (red). Note the almost perfect reproducibility, second order flux correlations between and the performance of two telescopes. It was shown that this Count ra 100000 Aqueye consistent with method allowed the coherence function that of a larger tele- of the emitting source to be measured, scope at a better site. 50000 with the advantage that the required tolerances on the wavefronts were centi- metres and millimetres, not microns 00.2 0.4 0.6 0.81.0 and nanometres as in Michelson interfer- Phase ometers. The disadvantage with intensity interferometry is that it requires very large The power of instruments such as Iqueye light curves of fast variables such as pul- photon fluxes, and thus has limited sensi- is only starting to become evident. The sars, but also to compare, on a precise tivity. This limitation can be significantly ability to measure events at nanosecond timescale, light curves taken at different overcome by the better detection efficien- timescales with accurate time-tagging times and at different locations. For cies and larger telescopes than were has already opened the door to a number pulsars, this enables the study of phase available in the 1960s; many in the com- of unique applications. It becomes possi- changes, derivatives, and comparison munity are debating the revival of inten- ble not only to measure the individual with radio observations, for example. It sity interferometry. In this context, plans to bring two Iqueyes to the VLT are IQuEYE−2009011 5−030304−CRAB2− already being proposed.

0.66 Autocor. To tal (coeff), ΔT = 0.0001s 0.64 In the future, with improved hardware and the immense photon collecting power of 0.62

n the European ELT, it should become 0.6 io possible to truly attain the quantum limit, at 0.58 el attempting detections at the picosecond 0.56

Corr level. This will open an entirely new win- 0.54 dow on the Universe, and push our knowl- 0.52 edge into the regime where even the con- 0.5 cept of photon and wavelength come 0.48 to lose their accustomed meaning: a truly 0.25 0.30.350.4 0.45 0.50.55 0.6 quantum eye will be born. Marinetti would (s) have loved it!

x 10 5 psrb0540−69 Light Curve − Bin time: 0.0025325 s Figure 7. (Top) A zoom 4.19 of the autocorrelation References function of single pulses from the Crab pulsar, Barbieri, C. et al. 2008, in Instrumentation for the 4.18 obtained by Iqueye at VLT in the ELT Era, ed. A. Moorwood, the NTT in January (Berlin: Springer), 249 2009. The structure in Dravins, D. et al. 2005, QuantEYE. Quantum Optics 4.17 the secondary peak is Instrumentation for Astronomy, OWL Instrument due to the double-pulse Concept Study, OWL-CSR-ESO-00000-0162 shape of the light curve. Fors O. et al. 2008, A&A, 480, 297 4.16 Redfern, M. & Ryan, O. 2006, presentation at the m e e t i n g Towards the European ELT, Marseille Richichi A. et al. 2006, The Messenger, 126, 24

tal counts 4.15 Richichi A. et al. 2008a, A&A, 489, 1399

To Richichi A., Fors O. & Mason E. 2008b, A&A, 489, 1441 4.14

Figure 8. (Left) The light curve of the pulsar 4.13 PSR 0540-69, with magnitude V = 23. The period can be well established from these data (preliminary value 50.7 ms), as well as features in the curve that 4.12 reproduce well previous observations by HST. The 00.010.020.030.040.050.060.070.080.090.1 data timing needs to be corrected for Earth’s rota- Time (s) tion, orbit and other effects.

36 The Messenger 135 – March 2009 Astronomical Science

VISIR Observations of Local Seyfert Nuclei and the Mid-infrared — Hard X-ray Correlation

Hannes Horst1 separated into two classes: Sy 1 galaxies the dynamical stability of the torus Poshak Gandhi2 with an unobscured view onto the hot, (Krolik & Begelman, 1988). Alain Smette3 optically bright accretion disc and the sur- Wolfgang Duschl1 4 rounding broad line region (BLR); and – Panel c illustrates the idea of the “reced- Sy 2 galaxies that are observed through a ing” torus, first proposed by Lawrence veil of gas and dust. This has led astrono- (1991). The consequence of this model is 1 Institut für Theoretische Physik und mers (see, for example, Rowan-Robinson, that AGN of high luminosity will appear Astrophysik, Universität Kiel, Germany 1977; Antonucci, 1982) to develop the as obscured less frequently than AGN of 2 RIKEN Cosmic Radiation Laboratory, unified scenario for AGN. The cornerstone low luminosity, since, when the luminos- Wako City, Saitama, Japan of this scenario is the existence of a torus- ity increases, the sublimation radius and 3 ESO shaped supply of molecular gas and thus the inner edge of the torus move 4 Steward Observatory, University of dust that obscures the central parts of the outwards. This, in turn, means that the Arizona, USA AGN when viewed close to the equatorial solid angle covered by the torus as seen plane (see Figure 1 for an illustration). from the centre of the AGN decreases. Thus, Sy 1 and Sy 2 galaxies are the same There are some observational indica- High angular resolution mid-infrared beasts, just seen from different directions. tions for this effect, but the question of observations with the VISIR instrument the geometry of the torus in this case is at the Very Large Telescope have While the unified scenario has proved to far from settled. allowed the distribution of dust around be very successful and has passed many local active galactic nuclei (AGN) to observational tests, little is known about There are many approaches to the study be studied. The observational results the physical state of the torus itself. Some of the physical state of the torus. The support the unified scenario for AGN possible geometries for the dust distribu- one that seemed most promising to us and bring constraints on the properties tion are shown in Figure 2: was to combine mid-infrared (MIR) and of its key component, a dusty torus hard X-ray observations. The intrinsic obscuring the view onto the AGN when – Panel a displays a classical smooth dis- X-ray luminosity is a good proxy for the viewed close to the equatorial plane. tribution with a constant ratio of height/ total luminosity of the accretion disc. The radius. Such a torus has essentially MIR emission of an AGN, on the other the same geometry as the accretion disc hand, is dominated by thermal emission Active galactic nuclei have been a prime in the centre of the AGN, but with much of the dust within the torus. As the torus target of extragalactic astronomy for lower temperatures and a much larger is heated by the accretion disc, one many years. These fascinating objects extent. The border between the torus on would expect a correlation between MIR comprise a supermassive black hole, a one side and the accretion disc and and hard X-ray luminosities. Since, in hot accretion disc feeding the black hole broad line region on the other side is Sy 1 galaxies, we can see the hot dust in and an additional supply of cold gas determined by the sublimation tempera- the inner part of the torus, while in Sy 2 and dust. They have many manifestations: ture of the dust. When the gas becomes galaxies — according to the unified sce- quasars, radio galaxies, Seyfert galaxies too hot, the dust particles are destroyed, nario — we can only see the cooler dust — all of which are governed by the same the gas is ionised and a hot BLR or in the outer part of the torus, one would physical mechanisms. accretion disc is observed instead of a expect to find a difference in the MIR and dusty torus. intrinsic hard X-ray luminosities.

The dusty torus in AGN – Panel b shows the same geometry, but Two studies on the mid-infrared–hard with the dust and gas arranged in dis- X-ray correlation by Krabbe et al. (2001) The local incarnations of AGN — also tinct clouds. There are theoretical argu- and Lutz et al. (2004) found the expected known as Seyfert (Sy) galaxies — can be ments for such clumpiness, based on correlation between MIR and hard X-ray

broad line region obscuring torus a) z b) z

θ h θ h r r

black hole accretion disc c) z

Figure 1. Sketch of the inner part of an AGN as pre- Figure 2. Vertical cuts showing three possible dicted by the unified scenario. Depicted are the geometrical configurations for the torus: a) shows black hole, the accretion disc, the clumpy broad line θ a “classical” torus; b) a clumpy torus; and c) a region and the obscuring torus. The narrow line h receding torus. r is the radius, h the height above region and a possible jet are omitted for clarity. r the equatorial plane and θ the half opening angle.

The Messenger 135 – March 2009 37 Astronomical Science Horst H. et al., VISIR Observations of Local Seyfert Nuclei

Figure 3. The central part of NGC 5135 in an overlay of optical Hubble Space Telescope data and our VISIR observations. The optical data is shown as a false colour image and the VISIR data as a contour plot. This figure is reproduced from Horst et al. (2009). luminosities, but not the expected differ- With this setup, we reached a typical reso- ence in luminosity ratio between Sy 1 and lution of 0.3 arcseconds, thereby signifi- Sy 2 galaxies. cantly improving the resolution compared to previous studies. Most of the observed Ramifications of assessing the MIR and objects appear point-like — only in three X-ray properties of AGN cases did we detect extended emission et

around the AGN. One example of a target fs One limitation of earlier studies on the with extended emission — NGC 5135 mid-IR–hard X-ray correlation was the low — is displayed in Figure 3. In this overlay angular resolution of the MIR observa- of VISIR and Hubble Space Telescope Dec Of tions. Unfortunately, Seyfert galaxies fre- data, we see that some sources of optical quently host regions of active star forma- and MIR emission coincide, while others tion close to the AGN which also emit in do not. Where MIR sources do not have 1. 0” the MIR regime. If it is not possible to spa- an optical counterpart, we probably see tially resolve these star-forming regions, the early stages of star formation when a separation, e.g., by spectral decomposi- young stars are heating their environment, RA Offset tion methods, is very difficult. We decided but are not yet powerful enough to blow to base our study on observations with away the dusty veil that is hiding them. contamination in our own data, but at the VSIR instrument (Lagage et al., 2004) least it will be far less than for larger at the Very Large Telescope (VLT), which aperture observations. We discuss the offers the best combination of spatial res- Extended emission and comparison to important point of contamination in our olution and sensitivity in the mid-infrared Spitzer data data in more detail in the next section. currently available in the world. For NGC 5135 and a second object, Besides the angular resolution of the MIR NGC 7469, we can estimate that within the The mid-infrared–hard X-ray correlation observations, another issue was the innermost 3.0 arcseconds around the reliability of our X-ray data. As mentioned AGN, at least roughly 45% of the MIR flux The primary goal of our observing cam- above, we need the intrinsic X-ray lumi- does not originate from the AGN. For paigns was to study the mid-infrared– nosity in order to have a proxy for the total NGC 7469, this result is in good agreement hard X-ray correlation. First of all, like our luminosity of the accretion disc. Unfor­ with an actual comparison of our VISIR predecessors, we found a strong and tunately, X-rays are also absorbed within data to archival spectra recorded with the significant correlation between these two the torus. However, high quality X-ray IRS instrument aboard the Spitzer Space quantities. The correlation is shown in spectra allow us to correct for this effect, Telescope; the latter provides an angular Figure 5. In this figure, we plot hard X-ray since absorption alters the X-ray spectral resolution of ~ 3.0 arcseconds. We per- v. MIR luminosity on a log–log scale. Blue ap­pearance of an AGN in a characteristic formed this comparison for all objects that squares represent Sy 1 nuclei, red dia- way. Therefore, we decided to observe we had observed in at least two MIR filters monds are for Sy 2 nuclei and green tri- a sample of Sy 1 and Sy 2 galaxies with and for which archival IRS data were avail- angles for low ionisation emission line VISIR for which high quality X-ray data able. Interestingly, we found some cases regions (LINERS), a class of AGN with were available. of significant deviation of the IRS and less pronounced nuclei and — relative to VISIR fluxes where no ex­tended emission the AGN power — more star formation In order to make the best use of VISIR’s was visible in the VISIR images (Horst et than Seyfert galaxies. Arrows mark either high angular resolution capabilities, we al., 2009). One example for this is shown the upper limits of the MIR luminosity decided to only observe relatively nearby in Figure 4. We interpret this discrepancy of non-detected sources or the lower lim- AGN. We set the limit for our sample as caused by smooth, extended emission its of sources with equivocal X-ray spec- selection at a redshift of 0.1. that is not observed with VISIR due to its tra. The dotted line depicts the best-fit limited sensitivity to this kind of emission. power-law to our first sample, observed In a few cases, flux changes due to time in 2005 (see Horst et al., 2006); the VISIR observing campaigns variability between the two observing dashed line then shows the best fit to the epochs, while unlikely to be the origin of two samples combined. Between April 2005 and September 2006, this effect, cannot be completely ruled out. we observed 29 Seyfert galaxies with The correlation has some important fea- VISIR and detected 25 of them. We used The extra-nuclear emission — either tures. One of these is the slope, which is a standard chop/nod-procedure to directly observed with VISIR or visible close to unity. On a log–log scale, a slope remove the very bright atmospheric back- through the comparison to the Spitzer of one indicates a linear dependency ground in the MIR. In order to achieve the data — underlines that giving priority to between the two plotted quantities. Thus, highest possible angular resolution, we high angular resolution, instead of the we find that MIR and X-ray luminosities chose VISIR’s small field objective, which higher sensitivity of space telescopes, show a linear relationship. Secondly, Sy 1 provides a pixel scale of 0.075 arcseconds was the right approach for our purpose. and Sy 2 galaxies exhibit the same MIR/ and a field of view of 19 x 19 arcseconds. Of course, we cannot rule out significant X-ray luminosity ratio. Again, this is in

38 The Messenger 135 – March 2009 Figure 5. The mid-infrared–hard X-ray correlation as determined in this study. Blue squares are Sy 1 nuclei, red diamonds are Sy 2 nuclei and green trian- gles are LINERS. Arrows either depict upper limits Figure 4. Comparison of VISIR photometry (black to the MIR luminosity, in the case of MIR non-detec- circles) with IRS spectrophotometry (black line) for tion, or lower limits to the X-ray luminosity, in case of Mrk 590. The horizontal error bars depict the pass- equivocal X-ray data. The dotted line is the best-fit band of the VISIR filters; dashed vertical lines indi- power-law to our first sample, the dashed line the cate the wavelength position of emission lines com- best-fit power-law to the combined sample. This fig- mon in AGN spectra. ure is reproduced from Horst et al. (2008).

250 Mrk 590 45 NGC7674 44 200

) PKS2048–57 –1

s 43

IRAS 13197–16 27 ) rg Jy

150 (e

v

(m MCG–01–01–043 ke 42

10 2– L Flux 10 0 41 NGC4579 log

40 50 Sy 1–1. 5 NGC4303 Sy 1. 8–2 39 LINERs NGC4698 0 8910 11 12 13 39 40 41 42 43 44 45 log L (12.3 µm) (erg s–1) Rest frame wavelength (µm) λ λ good agreement with the results of previ- Implications for the dusty torus torus-shaped distribution of molecular ous studies. At our high angular resolu- gas and dust, if the dust is arranged tion, it is unlikely that the similarity is Our results allow us to constrain the prop- in distinct clouds (panel b in Figure 2). In caused by contamination of the observed erties of the dusty torus — at least for addition, these clouds need to have a low MIR luminosity with non-AGN emission. the AGN within the luminosity range we volume filling factor within the torus. In Therefore, we assume that it is intrinsic to probed. Since we aimed for high angular such a configuration, we have a relatively AGN. resolution, we restricted ourselves to unobstructed view through the torus onto observing local AGN that are less lumi- the hot dust in its inner region. On account of the emphasis that we nous than more distant objects, e.g., qua- placed on the issue of angular resolution, sars. Thus, it has to be kept in mind that These interesting results have motivated we have to discuss the influence of con- the properties of the torus in AGN may us to continue research along these lines. tamination in our own data. In order to change toward higher luminosities (and in One important study was carried out obtain robust results, we split our sample fact there is some evidence that this is so). by Gandhi et al. (2009) who — thanks to into two sub samples: especially well- the arrival of the Suzaku, INTEGRAL and resolved objects and less well-resolved Interestingly, however, we find no indica- Swift spacecraft — managed to obtain objects. The term “well-resolved” has tion for a luminosity dependence of the reliable X-ray data for especially heavily to be understood in terms of the dust appearance of the torus within our sam- obscured Sy 2 galaxies. VISIR observa- sublimation radius that defines the inner ple. Since the slope of the correlation is tions of these targets showed that also edge of the torus. This radius can be esti- unity within the errors, the X-ray/MIR lumi- they follow the correlation found in our mated from the X-ray luminosity and is nosity ratio does not change at all. Since earlier studies. The implication is that our a natural scale for the dusty torus. Inter- the MIR luminosity is determined by the approach has indeed allowed us to con- estingly, we find a significant change in amount of accretion disc emission that is strain the geometry and physics of the the MIR/X-ray luminosity ratio at an angu- absorbed by the torus, it is directly pro- dusty torus in AGN. An interesting next lar resolution of 560 times the dust subli- portional to the hard X-ray luminosity and step would be to widen the luminosity mation radius (see Gandhi et al., 2009 for the solid angle it covers when seen from range covered by this study and try to details). Therefore, we used this resolution the accretion disc. Thus, a constant lumi- assess the properties of the tori in the as the separator for our two sub samples. nosity ratio implies that the opening angle least, as well as the most powerful, AGN. The well resolved AGN are marked by θ of the torus is constant as well. This black circles in Figure 5. We then checked rules out the receding torus model (panel whether the correlation would change c) in Figure 2 in its purest form. If the torus References if we used only the well-resolved objects; shows receding behaviour, it only does so Antonucci, R. 1982, Nature, 299, 605 38 reassuringly, the result of this exercise beyond X-ray luminosities of 10 W. Gandhi, P. et al. 2009, A&A accepted showed that this is not the case. Within Horst, et al. 2006, A&A, 457, L17 errors, the slopes of the two correlations The fact that Sy 1 and Sy 2 nuclei follow Horst, H. et al. 2008, A&A, 479, 389 are identical (the details of the statistical the same correlation also has an impor- Horst, H. et al. 2009, A&A, accepted Krabbe, A. et al. 2001, ApJ, 557, 626 analysis are presented in Horst et al., tant implication: it implies that the as­­ Krolik, J.H. & Begelman, C. 1988, ApJ, 329, 702 2008). Thus, we can be confident that the sumption that in Sy 1 nuclei we see hot Lagage, P.O. et al. 2004, The Messenger, 177, 12 correlation we have determined is physi- dust, while in Sy 2 galaxies we do not, Lawrence, A. 1991, MNRAS, 252, 586 cally meaningful and can now discuss its is probably incorrect. This result can only Lutz, D. et al. 2004, A&A, 418, 465 Rowan-Robinson, M. 1977, ApJ, 213, 635 implications. be reconciled with the existence of a

The Messenger 135 – March 2009 39 Astronomical Science

A VLT Large Programme to Study Galaxies at z ~ 2: GMASS — the Galaxy Mass Assembly Ultra-deep Spectroscopic Survey

Jaron Kurk1 200 spectra produced by GMASS con- which the most massive systems form Andrea Cimatti2 stitute a lasting legacy, populating the relatively late through a slow process of Emanuele Daddi3 “redshift desert” in GOODS-S. merging of smaller galaxies. Together Marco Mignoli4 with the small and slow evolution in the Micol Bolzonella4 K-selected galaxy population up to z of Lucia Pozzetti4 Motivation 1−1.5, it became clear that most (mas- Paolo Cassata5 sive) galaxy assembly occurred at z > 1.5. Claire Halliday6 In 2002, the K20 survey provided the In order to study the physical and evolu- Gianni Zamorani4 spectroscopic redshift distribution of a tionary status of typical Milky Way mass Stefano Berta7 complete sample of 480 galaxies with (M*) galaxies in this redshift range, we Marcella Brusa7 K < 20. One of the main scientific results proposed to obtain ultra-deep spectros- Mark Dickinson8 from this survey was the discovery of a copy with FORS2 at the VLT. As emission Alberto Franceschini9 significant population of massive lines move out of the optical window Giulia Rodighiero9 K-selected galaxies at high redshift and the redshift measurement, especially Piero Rosati10 (Cimatti et al., 2002). Their spectra for the passive galaxies, depends on Alvio Renzini11 (Cimatti et al., 2004) showed that some of absorption features in the continuum, these objects were indeed very massive there are few galaxies known in the 11 (10 M0) and already old (1–2 Gyr). redshift range of interest, 1.3 < z < 2.5, 1 Max-Planck-Institut für Astronomie, Around this time, deep near-infrared (NIR) which has traditionally been known as Heidelberg, Germany imaging surveys were also beginning the redshift desert. 2 Università di Bologna, Bologna, Italy to provide evidence of a new population 3 CEA/Saclay, Gif-sur-Yvette cedex, of candidate massive galaxies at photo- The location chosen for the survey (Fig- France metric redshifts z > 2 (Labbé et al., 2002), ure 1) was GOODS-S in the Chandra 4 INAF/Osservatorio Astrofisico di barely detectable even in the deepest Deep Field South, because of the avail­able Bologna, Italy optical images. The strong clustering of deep optical Advanced Camera for Sur- 5 University of Massachusetts, Amherst, these red galaxies (Daddi et al., 2003) veys (ACS) imaging and near-infrared USA suggested that these were progenitors imaging with the Very Large Telescope 6 INAF/OAA, Florence, Italy of local, massive early-type galaxies. The (VLT) ISAAC instrument. This field would 7 Max-Planck Institut für extraterres- existence of a significant population of also contain the planned Hubble Space trische Physik, Garching, Germany massive galaxies in the early Universe Telescope (HST) Ultra Deep Field (UDF) 8 National Optical Astronomy was not predicted by semi-analytic mod- and contained part of the K20 survey. Observatory, Tucson, USA els of hierarchical galaxy formation, in The proposed deep spectroscopy would 9 Università di Padova, Padova, Italy be complementary to the VIMOS and 10 ESO FORS2 public spectroscopy surveys car- 11 INAF/Osservatorio Astrofisico di ried out by ESO in the GOODS-S field, Padova, Italy which had shorter integration times and therefore targeted more luminous objects. GMASS Including overheads and pre-imaging, We report on the motivation, sample the total time requested for the programme selection and first results of our VLT was 145 hours in two semesters. FORS2 Large Programme (173.A-0687), which has obtained the longest targeted K20 spectra of distant galaxies obtained so UDF Sample selection and mask design far with the VLT. These long exposures, up to 77 hours for objects included For the target selection, we took advan- in three masks, were required to detect tage of the very recent Spitzer/IRAC spectral features of extremely faint gal- coverage of GOODS-S. The IRAC 4.5 μm axies, such as absorption lines of photometry, which samples the 1–2 μm passive galaxies at z > 1.4, a population rest-frame for the targeted redshift range, that had previously escaped attention enabled us to select on mass more relia- due to its faintness in the optical wave- bly than possible with the K-band. We length regime, but which represents combined all available photometry from a critical phase in the evolution of mas- GOODS-South U-band to 8.0 μm to obtain photometric sive galaxies. The ultra-deep spectros- HST + ACS (F850LP) redshifts for the 1277 unblended sources copy allowed us to estimate the stellar detected at 4.5 μm (to mAB = 23.0). These metallicity of star-forming galaxies at Figure 1. The location of the 6.8 x 6.8 arcminute field 1277 sources constitute the GMASS sam- z ~ 2, to trace colour bimodality up to of GMASS (red), relative to the ACS coverage of ple. Of this sample, we selected objects GOODS-S (black background), the field of the K20 z = 2 and to characterise a galaxy cluster survey (yellow) and the Hubble Ultra Deep Field for spectroscopy according to the follow- progenitor at z = 1.6. The approximately (green). ing constraints: zphot > 1.4 and BAB < 26

40 The Messenger 135 – March 2009 and IAB < 26, excluding all objects for Redshifts obtained 1.8, and is 100% for most bins in the which spectroscopy was already availa- range 1.8 < z < 2.9. ble or planned. This spectroscopic sample After extracting the one-dimensional contains 221 galaxies. Subsequently, spectra and fitting the absorption and we identified two subsamples: selecting emission line features, 130 new redshifts Superdense passive galaxies at z > 1.4 so-called blue galaxies that were ex­­ at z > 1.4 were obtained. In addition, pected to have strong absorption lines 37 new redshifts at z < 1.4 were obtained One of the main purposes of GMASS below 6000 Å for the masks to be and more than twenty formerly known was to discover (spectroscopically) and observed with the 300V grism (the blue redshifts were confirmed and, in most study passive galaxies in the redshift masks); and red galaxies to be observed cases, determined with higher accuracy. desert. Although this desert is no longer with the 300I grism (the red masks). For The fraction of redshifts successfully as empty as before, the galaxies known the first two pilot masks, we selected the determined for the targets observed is in this redshift range are mostly UV- brightest targets from both subsamples. about 85%. The same number applies to selected (Steidel et al., 2000) and there- Subsequently, we filled two blue and two the fraction of targets with photometric fore actively star-forming galaxies. Using red masks with fainter targets, allocating redshifts z > 1.4 that were confirmed by IRAC selection, we were able, instead, open spaces to targets already included in spectroscopy to have z > 1.4 (a few of to uncover 13 passive galaxies in the other masks, or otherwise to targets from the galaxies from the spectroscopic sam- range 1.39 < z < 1.99. Their spectra (see the GMASS sample without spectroscopic ple were included erroneously and are Figure 3) are similar to those of other old redshifts. In total, 211 (174) objects from not considered in this fraction). Figure 2 passive galaxies at z > 1, such as that of the GMASS (spectroscopic) sample were shows a histogram of the photometric 53w091 at z = 1.55 (Dunlop et al., 1996). included in one or more masks (38 objects and known spectroscopic redshifts in the were included in two masks and five GMASS field. Also shown is the fraction We obtained a high signal-to-noise in three). The total exposure times for the of known spectroscopic redshifts pro- stacked spectrum (shown in Figure 4) by masks were 12 h, 14 h and 15 h for the vided by GMASS, which increases from averaging all 13 individual spectra, as­­ blue and 15 h, 32 h and 30 h for the red 30% to 85% between redshifts 1.4 and signing the same weight to each spectrum masks.

Our observational strategy for the blue masks followed the conventional optical /DQBDMS@FDQDRTKSHMFEQNL&, 22 /GNSNLDSQHBQDCRGHES one of having two offset positions so 2ODBSQNRBNOHBQDCRGHES as to be able to correct for bad pixels, with 2ODBSQNRBNOHBQDCRGHESEQNL&, 22 each offset position exposed for half an  hour. For the red masks, for which many variable sky emission lines make back- ground subtraction difficult, we used four offset positions, each exposed for 15 minutes. During reduction, we com-  puted the median value of the four posi- tions for each pixel, which provided a reli- D able representation of the background. MS@F After subtraction of this background, the BD two-dimensional spectra were rectified  and combined (taking into account the respective offsets). The remaining sky-line residuals were fitted along the columns @MCODQ Y

and subtracted from the combined - image. The resulting spectra have very  low sky-line contamination, even above 8500 Å. We also took care to interpolate only once during the entire reduction process, to minimise any noise introduced by this process. 

Figure 2. Histogram of photometric (grey), previously known spectroscopic redhsifts (blue) and spectro- scopic redshifts resulting from GMASS (red) in  the GMASS field. Also shown is the percentage of known spectroscopic redshifts determined by  GMASS (black dots). /GNSNLDSQHBFQDXNQRODBSQNRBNOHBBNKNTQDCQDCRGHES

The Messenger 135 – March 2009 41 Astronomical Science Kurk J. et al., VLT Programme to Study Galaxies at z ~ 2: GMASS

 Figure 3. Individual spectra of six of the thirteen pas-    H    ! sive galaxies with GMASS spectra (blue). The spec-    tra are very similar to the overlayed spectrum (red) of 

  ( (( !  the old galaxy LBDS 53w091 at z = 1.55 (Dunlop et  (( ,F  ,F al., 1996). Shown on the right are 3.7 x 3.7 arcsec- . Y   ond BVI stamp images, constructed from HST/ACS H    ! images. 

 after normalisation in the 2600–3100 Å Y  l  wavelength range. The stacked spectrum

Ä H    ! l  was compared with various libraries of L synthetic spectra to estimate the age of

B 

l the stellar population. In practice, using Y   the observed rest-frame UV spectrum FR H   Q  ! only, provides an estimate of the time D   elapsed since the last major episode of l star formation. Since a degeneracy exists  Y  

h  between the effects of age and metallicity,

% H     ! we calculated ages for a range of metal­  licities, finding best-fit ages between 0.7  and 2.8 Gyr for metallicities Z = 1.5 to Y   0.2 Z . We also fitted synthetic spectra H   0   ! to the available photometry (11 bands, including HST/ACS BVIz, VLT/ISAAC JHK ,  s and Spitzer/IRAC 3.6, 4.5, 5.8, 8 μm) and Y   derived a mean age, mass, and upper limit        to the star formation rate (SFR) of 1.1 Gyr, Ä 10 hQDRS 5 x 10 M0, and 0.2 M0/yr respectively. Owing to the depth of the spectroscopy, it

 was possible to study galaxies that span (   Q a wide range in stellar mass: from very massive 1011 M galaxies to systems of 

 " 0    ((  '   10 

  D 

- only 10 M . Our analysis of the ages    0       ( ((   ((  indicate that the bulk of the stars in these  % (( (( ( + +

! ,M %D %D %D ,F ,F ! ! passively evolving galaxies must have formed at 2 < z < 3, which is in excellent 

 agreement with the evidence found for   early-types observed at 0 < z < 1, as well ! as for recent observations of early-types at z > 1.4.

A visual classification by eye of the ACS SX   images shows that the majority of passive galaxies have a spheroidal morphology typical of early-type galaxies (Figure 3).

%KTWCDMRH For data acquired with the reddest availa- ble ACS filter, we modelled the surface   brightness distribution by fitting a Sersíc profile, after the point spread function (PSF) was determined from ten stars in the field. Most of the passive galaxies have a Sersíc index n > 2, indicating that the   bulk of the light from the galaxy comes from a bulge component. The galaxy radii are in the range 0.6–3.2 kpc with a mean &, 22GNTQRODBSQTL

, &XQ22/ 599 Figure 4. The stacked average spectrum of the thir-   teen passive GMASS galaxies (black) overlayed with      the best-fit synthetic spectrum (red) from the models 1DRSEQ@LDV@UDKDMFSGÄ of Maraston (2005).

42 The Messenger 135 – March 2009         Figure 5. Density maps of galaxies at z = 1.6 in the GOODS-S field. Filled contours are based on GOODS/MUSIC (Grazian et al., 2006) photometric t redshifts (indicated by filled circles) and solid and dashed contours are based on spectroscopic red- $KKHOSHB@K shifts from GMASS and ESO/GOODS (indicated by 2OHQ@K crosses). Small dots indicate galaxies in the GMASS catalogue outside the redshift spike. Additional sym- t (QQDFTK@Q bols are plotted according to morphology: elliptical -NSHM&, 22R@LOKD galaxies (circles), spiral galaxies (squares) or irregu- lars (triangles). Diamonds indicate those galaxies in the overdensity at z = 1.6 not in the GMASS sample. t leads to an estimated sub-mm/mm gal- axy duty cycle of ~ 0.15 Gyr. SMGs may therefore represent rapid and highly t dissipative major mergers at z > 2, which  become compact, superdense remnants that evolve almost passively at z < 2.

)  t M The other question is how superdense N SH &, 22%.12%HDKC galaxies at z ~ 2 migrated to the size- mass relation at z ~ 0. One possibility is t dissipationless (or dry) merging: accord- #DBKHM@ ing to some models, this process can increase the size and mass of a system t without altering the stellar population content. The increase in size is expected to depend on the orbital properties and the mass of the merging system, with the t most massive galaxies increasing their size more strongly than their less massive counterparts. If the mass-dependent size evolution is applicable, most GMASS t - passive galaxies would be the progeni- tors of early-types that today have stellar 6 11 12 masses 10 M0 < M < 10 M0 , i.e. the Œt most massive E/S0 systems at z = 0. GLR LR  R R R  R R 1HFGS RBDMRHNM) An overdensity of galaxies at z = 1.6 of 1.4 kpc. These values are much smaller This significant difference in size raises Several spikes in the histogram of galaxy than those observed in early-type galaxies two questions: firstly, how did these small redshifts in the GOODS-S field are known. of the same stellar mass in the local systems form, and what mechanisms The overdensity of one of the highest Universe. The small sizes that we measure can explain their growth in size in the redshift spikes, at z = 1.6, was described are neither due to an unresolved central past 10 Gyr? The derived constraints on in Castellano et al. (2007). Within the source, nor due to our imaging being the age, star formation history, and stellar large-scale overdensity, a peak was dis- in the rest-frame mid-UV (several authors masses indicate that intense star for­ cerned that would evolve into a cluster have demonstrated that the sizes of sphe- mation (SFR > 100 M0/yr) must have of galaxies, although from the evidence roidal galaxies do not vary substantially as taken place at z > 2. Among the possible at z = 1.6 the structure is unlikely to be a function of wavelength). The measured precursor candidates, only sub-mm/mm- virialised. Since the peak is located in our galaxy sizes are smaller by a factor of selected galaxies (SMGs) have sizes field, GMASS added 32 galaxies with two to three compared with z ~ 0 galaxies, and mass surface densities comparable confirmed z = 1.6 redshifts to the ten pre- implying that the stellar mass surface den- to those of the passive galaxies at viously known (from ESO/GOODS spec- sity of passive galaxies at < z > ~ 1.6 is 1.4 < z < 2. In addition, the correlation troscopy). We confirm that there is a five to ten times higher. Such superdense lengths and estimated masses of these significant, narrow spike in the distribution early-type galaxies with radii > 1 kpc are two populations are similar. SMGs are, of spectroscopic redshifts at z = 1.610, extremely rare in the local Universe (Shen however, an order of magnitude rarer which forms an overdensity in redshift et al., 2003). than these passive descendants, which space by a factor of six. The velocity

The Messenger 135 – March 2009 43 Astronomical Science Kurk J. et al., VLT Programme to Study Galaxies at z ~ 2: GMASS

Figure 6. The average combined spectrum of 75  

 star-forming GMASS galaxy spectra. Lines indicate    interstellar rest-frame mid-UV absorption features   (dotted), photospheric absorption features (dot–      short dash), and C iii 1909 Å emission (dot–long       

 dash).                                                                         h h hh   hh     (  h      h       ((   (( ( (  (( (5 (( (( (( (5 (5 ((( (( (( (( ( (( (5 ((( (( (( 5

9M 2H 2H 2H 2 . 2H 2H 2H 2H " " " %D -H ,F -H -H -H published. In Halliday et al. (2008), we Wl study the stellar metallicity of galaxies at z ~ 2, using a stacked spectrum of 75 star-forming galaxies (see Figure 6), cor-

h responding to a total integration time Wl 1HWDS@K

X% of 1653 hours. With this spectrum, it was  ÄHMCDW possible for the first time to measure

DMRHS the iron abundance and thus the stellar l W metallicity at a median redshift of z = 1.88. We constrain the metallicity to %KTWHMS log(Z/Z0) = –0.57 ± 0.16. Cassata et al. Wl (2008) describe the colour bimodality in the galaxy population up to very high redshifts. We find that the red sequence Wl of passive galaxies is recognisable up          to z ~ 2, but then disappears. In addition, 1DRS EQ@LDV@UDKDMFSGÄ Daddi et al. (2007a,b) used GMASS spec- tra, in combination with other spectra of distant galaxies, to study star forma- dispersion of these 42 galaxies is 450 redshifts to be compiled using the J-K tion and obscured active galactic nuclei km/s, which should increase with redshift colour, which brackets the 4000 Å break (AGN). Several more papers are in prepa- to become comparable with that of a at this redshift. We detect a red sequence, ration, describing the survey strategy cluster of galaxies. The redshift distribu- which is consistent with a theoretical and resulting redshifts (Kurk et al.), the tion is not Gaussian, but rather bimodal sequence of galaxies that formed their morphological analysis (Cassata et al.) with a primary peak at z = 1.610 and a stars in a short burst at z = 3. and dust properties of galaxies up to secondary at z = 1.602. Although we do z ~ 2.5. The published GMASS spectra not detect significant spatial separation This is the first and only structure of this will be publicly available through the ESO between the primary and secondary nature known: its redshift is higher than GOODS website, forming a lasting legacy peaks, this may be additional evidence that of any known galaxy cluster and for studies of high redshift galaxies. that the structure is not yet virialised. the structure contains spectroscopically Towards the northern part of the GMASS confirmed, red, early-type galaxies. Its field, the surface density of spike galaxies irregularity in angular and redshift space, References is five times higher than in the remainder including at least two localised higher Cassata, P. et al. 2008, A&A, 483, L39 of the field. The properties of spike gal­ density peaks, suggests that the struc- Castellano, M. et al. 2007, ApJ, 671, 1497 axies in this high density region are differ- ture is still relaxing, which is consistent Cimatti, A. et al. 2002, A&A, 391, L1 ent from those outside: the mean age with its low X-ray emission. Since this Cimatti, A. et al. 2004, Messenger, 118, 51 and mass are higher, while the mean star structure exhibits some of the properties Cimatti, A. et al. 2008, A&A, 482, 21 Daddi, E. et al. 2003, ApJ, 588, 50 formation rate (SFR) and specific SFR typical of clusters, it may be the progeni- Daddi, E. et al. 2007a, ApJ, 670, 156 (SFR per unit mass) are lower. Six out of tor of a cluster of galaxies, observed at Daddi, E. et al. 2007b, ApJ, 670, 173 the eight passive galaxies in the spike are its assembly. Dunlop, J. et al. 1996, Nature, 381, 581 in this region, three of which appear to Grazian, A. et al. 2006, A&A, 449, 951 Halliday, C. et al. 2008, A&A, 479, 417 form a sub-group of size smaller than Labbé, I. et al. 2002, The Messenger, 110, 38 90 kpc. This group of passive galaxies is Other results from GMASS Maraston, C. 2005, MNRAS, 362, 799 not located at the peak of the surface Shen, S. et al. 2003, MNRAS, 343, 978 density of spike galaxies (Figure 5), but Apart from the two projects described Steidel, C. C. et al. 2000, ApJ, 604, 534 about 800 kpc away. The available above, published in Cimatti et al. (2008) NIR imaging allows a colour-magnitude and Kurk et al. (submitted), two other diagram of the galaxies with confirmed papers based on GMASS spectra are

44 The Messenger 135 – March 2009 Credit: IAU/Lee Pullen Credit: IAU/Jose Francisco Salgado playing to delegates. to playing Quartet Kronos the Lower: speakers. of panel distinguished the with seen is ceremonies, of master Jarre, Jean-Michel Upper: Paris. in ters headquar UNESCO the at held 2009 Astronomy of Year International the of opening official the of Twoviews - Astronomical News Astronomical News

VirGO: A Visual Browser for the ESO Science Archive Facility

Evanthia Hatziminaoglou1 VirGO2 offers an alternative to this “tradi- 2007, 2008), whose target audience is Fabien Chéreau1 tional” querying form, by providing a the general public, and it therefore bene- visual impression of the available data, fits from many standard features such their overlaps and the instrument foot- as the display of star and planet posi- 1 ESO prints, while allowing the resulting frames tions, landscape rendering, intuitive real- to be filtered in real-time, before launch- time navigation, a variety of projection ing the query to the ESO archive. This modes, etc. To this set of standard fea- VirGO is the next generation Visual solution presents many challenges in the tures, VirGO adds the necessary tools for Browser for the ESO Science Archive field of user interface design, such as browsing through the ESO data archive. Facility (SAF) developed in the Virtual displaying and navigating through many Figure 1 gives an overview of the general Observatory Project Office. VirGO observations simultaneously without con- look-and-feel of VirGO. The main features enables astronomers to discover and fusing the observer, or filtering out and of VirGO are: select data easily from millions of selecting relevant observations in an intu- observations in a visual and intuitive itive way. An important aspect of VirGO 1) The main Graphical Window contains way. It allows real-time access and the is its capacity to access and handle large the dynamical view of the observations graphical display of a large number data collections on both the client and in the current field of view. Images of observations by showing instrumen- server sides, as well as to exchange data with footprints and previews (Rite et al., tal footprints and image previews, as with other VO tools. These two features 2008) as well as superimposed spectra well as their selection and filtering are achieved thanks to VO standards: the can be visualised simultaneously, all for subsequent download from the ESO Simple Image Access/Simple Spectral on a multi-resolution Digitized Sky Sur- SAF web interface. It also permits the Access (SIA/SSA) protocols for retrieving vey (DSS) background, if the user so loading of external FITS files or VOTa- images and spectra, respectively, from desires (see below). bles, as well as the superposition of a variety of astronomical repositories Digitized Sky Survey images to be used through a uniform interface; and the 2) The List Browser displays a single sum- as background. All data interfaces are PLASTIC (Platform for Astronomy Tool mary line for each frame selected in based on Virtual Observatory (VO) InterConnection) protocol for communi- the Graphical Window. The displayed standards that allow access to images cation with other VO tools. information includes observation date, and spectra from external data centres, exposure time, filter and instrument. and interaction with the ESO SAF web The existence of a preview is also indi- interface or any other VO applications. Overview of main features cated by the presence of a tick-box.

VirGO is a plug-in for the open source 3) The Info Window contains detailed VirGO as an alternative to the traditional (GPL) software Stellarium (Chéreau et al., information about the observation ESO archive query

The main ESO archive query form1, a web interface, allows the user to search the ESO archive based on Target Name 1 or Coordinates, Observing Date or Pro- 2 gramme ID, request the type of observa- tions (Imaging, Spectroscopy, Interfer­ ometry etc.) and the Category (Science or Calibration frames). The query may 3 take anything from a few seconds to sev- eral tens of minutes, depending on the volume of the requested data required by the user. The information is returned in the form of a table, listing the frames, as well as the position, programme ID, expo- 4 sure time, filter and other quantities and is followed by two tables summarising the total number of frames and exposure times, broken down by the various instru- ments or exposure time/instrument/filter.

5 Figure 1. Overview of the main Virgo window. For an 7 6 explanation of the numbered items, see text.

46 The Messenger 135 – March 2009 currently selected in the List Browser. 7) Finally, the Menu Bar gives quick Figure 2. An example of the multi-resolution DSS It also provides a direct link to access access to functionalities such as the background implemented in VirGO. the datasets, the preview images or download of observations from the transmission curves if available. ESO SIA/SSA servers; the activation of visualisation of images, superimposition the DSS background, grid display, etc. of catalogues, or access of related data 4) The View Selector allows the user and information from e.g., Simbad or to choose which observations to show/ In the current release of VirGO (1.4), VizieR), SPLAT-VO5 (a spectral analysis hide by defining a set of constraints observational data are accessed from tool) or TOPCAT6 (an interactive graphical such as observation type (images ESO Science Archive Facility servers viewer and editor for tabular data). For or spectra), processing type (raw data, using VO SIA/SSA services. The VOTables example, FITS images can be sent to Ala- highly processed data etc.), date or provided by these services are loaded din for quick visualisation before being exposure time. The second part of this by VirGO in streaming mode to allow a requested from the archive. Similarly, all window is the Tree Browser, which fluid interaction even when downloading reduced spectra available as Advanced contains a telescope/instrument/filter large datasets. The background is a multi- Data Products (comprising most UVES tree reflecting which observations are resolution false-colour JPEG version and HARPS spectra as well as FORS2 currently loaded in VirGO. It gives of the DSS originally created at STScI3 by and VIMOS GOODS spectra) can be visu- refined control over what to show/hide processing and combining the original alised using SPLAT-VO. VirGO’s commu- from the Graphical Window. blue, red and near-IR DSS images, allow- nication with the VO world is a two-way ing for more flexibility, e.g., at the time communication. For instance, it can 5) The Target Selection uses either the of defining an archive query. The special receive and display catalogues retrieved Simbad name resolver or the exact version hosted at ESO was post-proc- using Aladin or TOPCAT. coordinates. essed and indexed for use in VirGO (see Figure 2 for an example). 6) Additional functionalities in form of Use cases Tabs allow the user to see the servers being queried, select grouping options Interaction with VO tools and services We illustrate two examples of where using (e.g., footprint blending) and frame VirGO has advantages over the archive blending (depending on the number of VirGO is able to communicate through query form. superposed frames), or to send a the PLASTIC protocol with a variety of direct query for the selected frames to other VO-compliant tools and services, – A user preparing an observing proposal the ESO Archive. such as Aladin4 (a sky atlas allowing for in the Chandra Deep Field South (CDF-S)

Figure 3. Example out- put of the Advanced Data Products query form for ESO spectros- copy in the Chandra Deep Field – South (CDF–S).

The Messenger 135 – March 2009 47 Astronomical News Hatziminaoglou E., Chéreau F., VirGO: A Visual Browser for the ESO Archive

Figure 4. Output of VirGO from a search for spectroscopy in the CDF-S. Each object within a box has a spectrum, and zoom-in and visualisation of a selected spec- trum using SPLAT-VO is shown. )

/Å GDS J033226.84-274818 .9 versus Wavelength /s

2 0.03 cm

g/ 0.025 er 6

—1 0.02 0 (1

.9 0.015 18

48 0.01 -27 0.005 84

226. 0

–0.005

6500 7000 7500 8000 8500 9000 9500 GDS J033 Wavelength (Å)

wants to know about all the available or even visualising individual spectra – on the technical side, the immediate ESO observations within one square using other VO tools (here SPLAT-VO) objectives include the improvement of degree in this field. Using the main ESO before requesting them from the server scalability, the optimisation archive query form, the user will have to archive. of image and catalogue access and the wait about 28 minutes for all the frames use of improved caching, as well as (more than 31000) to be returned in a deeper integration with the ESO web- the form of an HTML table. The same Future developments site. query using VirGO will require only four minutes and will return the list in a Although fully functional, VirGO is a recent visual form, like that shown in Figure 1, application, and several major capabilities References allowing the user to see the overlaps as well as technical features still have to Chéreau et al. 2007, Stellarium 0.10.0, between observations immediately, be implemented. The next objectives are www.stellarium.org to get a feeling for the depth of the the following: Chéreau F. 2008, in ASP Conf. Series, Vol. 394, exposures (using the footprint blending ADASS XVII, ed. Argyle R. W., Bunclark P. S. & Lewis option) or to look selectively at the – add access to more ESO data, espe- J. R., 221 Kapadia et al. 2007, in Christensen, L.L. & Zoulias, reduced data. cially science-ready data such as Large M. (eds.) Communicating Astronomy with the Programmes and Surveys, large sets Public, 2007 – A user wants to know which targets of pipeline-processed data, and Press Rité, C. et al. 2008, in ASP Conf. Series, Vol.394, have been followed up spectroscopically Release images; ADASS XVII, ed. Argyle R. W., Bunclark P. S. & Lewis J. R., 605 by ESO/GOODS in the CDF-S. The Advanced Data Products query form7 – dissemination: some features of VirGO will again return an HTML list of files will be included in the standard version Notes with the relevant information, like that of Stellarium, e.g., the display of Press 1 http://archive.eso.org/eso/eso_archive_main.html shown in Figure 3. The same query Release images (Kapadia et al., 2007). 2 http://archive.eso.org/cms/virgo using VirGO will return a visualisation VirGO/Stellarium, along with other VO 3 http://www-gsss.stsci.edu/Acknowledgements/ of the available reduced spectra, shown tools, is also brought to schools in DataCopyrights.htm 4 in Figure 4. The user can then select organised practical sessions in class- http://aladin.u-strasbg.fr/ 5 http://astro.dur.ac.uk/~pdraper/splat/splat-vo/splat- an image to use as a background to rooms with teachers and students, as vo.html identify the sources or, by zooming in, part of a dedicated outreach effort within 6 http://www.star.bris.ac.uk/~mbt/topcat/ see the orientation of the slit (whenever AIDA8 (Astronomical Infrastructure for 7 http://archive.eso.org/eso/eso_archive_adp.html 8 available) or the number of spectra Data Access) and this effort will continue http://cds.u-strasbg.fr/twikiAIDA/bin/view/ EuroVOAIDA/WebHome composing the final product per object, in the future;

48 The Messenger 135 – March 2009 Astronomical News

News from the ESO Science Archive Facility

Nausicaa Delmotte1 Figure 1. A view of part (for the ESO archive team) of the ESO Data Centre where the ESO Science Archive is stored and accessed. 1 ESO

The latest developments from the ESO archive are presented. Information is provided to the astronomical commu- nity on new data releases and services.

Blu-ray discs for data distribution

The new Blu-ray disc technology will be introduced mid-2009 in the Science Archive Facility. Besides FTP transfer for small volumes of data, the archive is cur- rently offering its users the following media for data distribution: CD, DVD-R, and USB discs for data requests larger on a free-standing 6-metre high tower introduction of the automatic data trans- than 60 GB. The reasons for the change located about 50 m west of the APEX fer through the network from Paranal are the adoption of the most economical telescope and data are recorded every to Garching (Zampieri et al., 2008) has technologies and the adaption to larger minute. APEX weather data, beginning led to significantly improved data delivery data volumes. A Blu-ray disc has about 1 January 2007, are available from the services. PIs, can now access their own ten times the capacity of a single-layer archive2 and can be searched by date raw data, in most cases only a few hours DVD. With the introduction of the Blu-ray interval, or any meteorological parameter after the observations have completed. discs, the archive will stop offering CDs such as temperature, humidity, wind as an archive data distribution medium. speed/direction or precipitable water Therefore, archive users anticipating large vapour. Contact data requests and Principal Investigators (PIs) expecting significant data deliveries For more information about the ESO from their observing programmes are New data releases archive, or to subscribe to the archive advised to make sure that they have Blu- RSS feed to be informed about the latest ray disc readers available. Several major scientific data releases archive developments, see the archive have taken place through the ESO web page3. archive over the last months and are New query forms summarised here. The 30 Doradus raw For any questions or comments on the data taken during the HAWK-I commis- ESO archive, contact us at archive@eso. Besides the general archive query form sioning phase were released in Decem- org. that gives unified access to the com­plete ber 2008 and the raw data from the ESO collection of raw data, the Science AMBER/FINITO/UTs science verification Archive Facility also offers several instru- were released in October 2008. The References ment specific and technically oriented second release (DR2) of advanced data Lilly, S. et al. 2008, The Messenger, 134, 35 access points for astronomers who are products from the zCOSMOS redshift Zampieri, S. et al. 2008, to appear in ASP Conf. already familiar with ESO instrument set- survey (Lilly et al., 2008) took place in Proc. ADASS XVIII ups and observing strategies. Two new October 2008. It contains the results of instrument specific query forms1 have the zCOSMOS-bright spectroscopic Notes been released: one for AMBER (near- observations that were carried out with infrared instrument of the VLT Interfer­ VIMOS in service mode during the period 1 http://archive.eso.org/cms/eso-data/instrument- ometer) raw data and one for HAWK-I April 2005 to June 2006. specific-query-forms 2 (near-infrared wide field imager) raw data. http://archive.eso.org/wdb/wdb/eso/meteo_apex/ form 3 http://archive.eso.org/ The database of ambient conditions for Faster access to raw proprietary data the Paranal and La Silla observatory sites has been extended to include weather The modification of the archive systems information from the APEX observatory and interfaces to allow PIs to request site. The APEX weather station is located their own raw proprietary data and the

The Messenger 135 – March 2009 49 Astronomical News

New Infrastructures Require New Training: The Example of the Very Large Telescope Interferometer Schools

Paulo Garcia1 utmost importance as new facilities come European training in optical interferometry. (on behalf of the European Interferometry online. This article shares the very suc- One goal of the schools was to educate Initiative) cessful experiences of the VLTI training a new generation of young astronomers, schools project ONTHEFRINGE that equipping them with the ability to carry took place between January 2006 and out scientific programmes at the VLTI 1 Universidade do Porto, Portugal December 2008. (from preparation to data reduction and analysis). Another important goal was to place optical interferometry in context The discovery space for astronomy is The birth of the project with other techniques in key astronomical dramatically widening with the opera- areas of European leadership such as tion, construction and planning of At the start of the training schools project, adaptive optics and radio/sub-mm inter- ambitious new infrastructures. A key the VLTI was operational, but had yet to ferometry. In contrast with the NEON aspect of the scientific return from these ramp up and become the top optical observing schools that took place at facilities is the training of its users. We interferometric facility in the world. Exper- observatories, the VLTI schools were held report on a series of summer schools tise in optical interferometry was concen- in relatively geographically isolated loca- designed to train a new generation of trated in a few institutes involved in instru- tions, but with sufficient computing young astronomers in optical interfer- ment building and was not widespread capacity for hands-on training. The very ometry with the Very Large Telescope across Europe and the ESO user commu- nature of the VLTI made it unreasonable Interferometer. nity. If this uneven distribution of expertise to carry out the training at the La Silla continued it would clearly inhibit the sci- Paranal observatory. Since VLTI observa- entific maximisation of the investment in tions are normally carried out in service Eleven years have passed since first light the new interferometric infrastructure. mode, its location in Chile and the full-time on the Very Large Telescope (VLT). science use of the facility meant that the The last decade has been a vibrant one A few schools had been organised previ- schools obviously could not take place for astronomy. Common key words in ously with FP5 funding (for example, on-site. contemporary astronomy such as “dark the Les Houches school in 2002) and energy” or “exoplanets” appeared for institute funding (for example, the Leiden The project consisted of four schools: the first time in the title of a refereed jour- schools in 2000 and 2004), but any wider nal article only about ten years ago. The coordination was lacking. In contrast, – Observation and Data Reduction with thrust of astronomical discovery is driven a very successful annual programme of the Very Large Telescope Interferometer, by carefully planned new facilities and optical interferometry summer schools Les Houches, June 2006 infrastructure. The last decade has wit- (the Michelson/Sagan Summer Schools) nessed the deployment of the VLT and the has been running in the US since 1999. – Circumstellar Disks and Planets at Very VLT Interferometer (VLTI), the planning and The creation of the European Interferom- High Angular Resolution, Porto, May– initial construction of the Atacama Large etry Initiative (EII) network signalled the June 2007 Millimeter/submillimeter Array (ALMA) and beginning of a European-wide coopera- the planning of the European Extremely tion between countries with and without – Active Galactic Nuclei at the Highest Large Telescope (E-ELT). These new expertise in optical interferometry. Under Angular Resolution: Theory and Obser- facilities will come online in the next dec- the auspices of EII, the ONTHEFRINGE vations, Torún, August – September ade. The E-ELT will enable tremendous project was submitted to the European 2007 gains in sensitivity, making it possible, for Commission for FP6 and awarded fund- example, to probe the acceleration of ing. The project was coordinated by – Astrometry and Imaging with the the Universe. The profound gain of the Universidade do Porto/CAUP, with ESO, Very Large Telescope Interferometer, ALMA interferometer in both resolving Observatoire de Paris/LESIA, Max- Keszthely, June 2008. power and sensitivity is driven by three Planck-Institute for Astronomy, Heidel- key science goals, one of them being the berg and INAF/ Osservatorio Astrofisico The first and fourth schools were data study of the physics and chemistry of di Arcetri as partners and, as third reduction schools, aimed at hands-on planet-forming discs around young stars. parties, the Laboratoire d’Astrophysique observation preparation and data reduc- de Grenoble, the Nicolaus Copernicus tion, the first focusing on AMBER/MIDI New opportunities come with new chal- University in Torun, Poland and the and the fourth on PRIMA and image re­­ lenges and these were clearly identified Konkoly Observatory of the Hungarian construction. The second and third by the ASTRONET Infrastructure Road- Academy of Sciences. schools were science schools aimed map (Bode et al., 2008): “Recruiting and at placing optical interferometry in a training the future generation of Europe- wider context, in two fields where it has ans with advanced scientific and techno- The goals of the schools a major impact. logical skills is therefore a key aspect of any realistic Roadmap for the future.” The ONTHEFRINGE project aimed to Although the schools were open to all Training new generations of astronomers overcome the training gap by providing an researchers, EU requirements only on new observational techniques is of the integrated and structured approach to allowed funding to particular categories

50 The Messenger 135 – March 2009 of participants, essentially PhD students used to constrain models of the astro- Highlights and results of the schools and young postdocs. nomical object(s) to be observed. Then, more complex and realistic aspects of The total number of participants in the the observations were included, such as four schools, including lecturers, was The design of the schools UV coverage, closure phases, noise in around 280. A lecturer attending four the observables and model-fitting. In the schools counted as four participants. Previous experience with the one week data reduction schools, the participants The number of participants who did not school at Les Houches in 2002 showed went through the steps in the reduction lecture was around 200, composed that it was difficult to combine both the pipelines. At the end of the schools, essentially of PhD students and young teaching of new information and the groups of students worked on a VLTI postdocs. The selection procedure kept necessary contact-building time among observation proposal and presented it multiple attendance to below 10%. Dur- participants successfully, in such a short to their colleagues, facing scrutiny and ing the three years of the project there time. Although, in theory, optical interfer- advice from seasoned observers. These were about 1500 PhD students in astron- ometry is not very difficult to understand sessions helped to mature a student’s omy, about half of whom were in Europe (the practice, however, requiring a signifi- understanding of the technique and cer- (Gibson, 2002). The total number of ESO cantly steeper learning curve), most of tainly helped to make them more able to member country participants was around the students had not been in contact with design scientific programmes at the VLTI 105. If we assume that half of the total the discipline before, in contrast with opti- – the main goal of the schools. PhDs are awarded in ESO member coun- cal/infrared spectroscopy or photometry, tries, then this project reached around for example. Therefore they had to be As a requirement of an EU funded train- 14% of the total PhD student population immersed in a Fourier space environment ing programme, we designed a series in Europe. This number would increase for some time to become acquainted with of lectures on complementary skills — by a small factor if specific areas were optical interferometry methods. The from presentation skills, to paper and tel- considered, such as ground-based schools lasted for two weeks, a couple escope proposal writing, career devel­ observational astrophysics. The gender of days more than the ideal. We found opment and ethics. We were surprised at balance of the non-lecturing participants that including a free afternoon as early the interest of the students. Many were was around 60% male: 40% female and as the second day of the school greatly not aware of what an impact factor was, increasing to 70% male: 30% female if improved the contact-building among or what the refereeing process was all we included the lecturing participants. participants. The right balance of fun and about, or even that job hunting should Figure 1 presents the distribution of non- work was found important to keep pro- start well in advance of the end of a PhD. lecturing students by country. Interest- ductivity high in the unfamiliar environ- Many of them were discussing ethical ingly, countries that have recently become ment of optical interferometry. It should aspects in the practice of astronomy members of ESO or intend to join soon be stressed that during fun time, students research for the first time, or learning that had a high attendance at these schools. were building contacts among themselves a PhD is not enough (c.f. Feibelman, and the senior lecturers that not only were 1993) for a successful career in science. These schools were very important for the basis for group work at the school, software developers (at ESO and the but could, in the future, be important for The number of students attending each Jean-Marie Mariotti Center) and they pro- collaborations or the sharing of informa- school was around 55, with PhD students vided a unique opportunity for immediate tion about relevant institutes and supervi- accounting for well over two thirds of and massive feedback from a pool of sors for a future postdoctoral position. the total. Such a number was a compro- interested new users. mise between a school environment and The school’s time was essentially divided reaching a larger audience. The logistics The ambiance of the schools was uni- between lectures/seminars and practical required to keep 30 computers up and formly excellent, providing a perfect bal- sessions. The lectures addressed the running (including updating and installing ance between hard-working and brain- basic aspects of interferometry and the software and data) with internet access cooling moments. The best experts and seminars the typical astronomical results in isolated regions of Europe was not triv- seasoned lecturers taught not only all obtained. In the astrophysical schools a ial. Even with so many participants, many about optical interferometry, adaptive set of review lectures focused on the more had applied and could not be optics, radio/sub-mm interferometry, but astrophysics of the target objects and selected to attend. The selection of partic- also the physics of young stellar discs or complementary techniques, such as ipants was based on a motivation letter, active galactic nuclei. The participants, adaptive optics or radio/sub-mm interfer- institute and national balance, with prefer- from all around the globe, had a taste of ometry. In the practical sessions the stu- ence given to PhD students. In order to the best of Europe! Sampling red wine dents went through the observation prep- reach a wider audience, all the materials and French cuisine at the Chateau de aration software, in groups of two sharing at the schools (presentations, software, Goutelas, or enjoying the sun and the a computer. This experience was found data, lecture notes) were made available beach at the Portuguese seaside (and of to be very useful in learning the basics of on the project website1. In retrospect course a glass of Port), visiting the stun- interferometry. Exercises focused initially some of the lectures should have been ning town where Copernicus was born on the observables measured by the videotaped and streamed through the site. and becoming an expert in wódka, or interferometer and how they could be finally watching the views from the hills

The Messenger 135 – March 2009 51 Astronomical News Garcia P., Example of the Very Large Telescope Interferometer Schools

Students not lecturing Acknowledgements This project was supported by EC contract No. MSCF-CT-2005-029954. We warmly thank the Italy consortium institutions and staff (Universidade do Porto/CAUP, ESO, Observatoire de Paris, MPIA, Germany INAF-Arcetri, Jean-Marie Mariotti Center, Torun Spain University and Hungarian Academy of Sciences) for France their full support. Special thanks go to the lecturers Finland + Czech Republic + Austria and students; kudos too to Júlio Carreira, Gilles Belgium Duvert, Manuel Monteiro and Elsa Silva. Remaining ESO members Poland References Russia Hungary Bode, M. F., Cruz, M. J., Molster, F. J. (eds) 2008, Remaining European countries The ASTRONET Infrastructure Roadmap, ISBN: 978-3-923524-63-1 Latin America Feibelman, P.J. 1993, A PhD Is Not Enough: A Guide USA + Canada To Survival In Science, (Basic Books: New York), Other countries ISBN: 978-0201626636 Gibson, B.K. 2002, Astr. Soc. of Australia Newsletter, 26, 4 Figure 1. Pie chart showing the distribution of stu- room full of students working hard on dents (non-lecturing participants) by host country their telescope proposals. At the end of Notes attending the four VLTI Summer Schools. the telescope proposal presentations, 1 http://www.vlti.org the students would rush to Françoise Delplanke and proudly collect ESO stick- near Lake Balaton while discovering the ers, posters, hats or calendars. This new bouquet of an old Tokay. I will not forget generation of young astronomers surely Figure 2. A selection of group photographs from the four VLTI summer schools at (clockwise): Les the awed face of lecturers as, when deserves the great new infrastructure we Houches, France, 2006; Porto, Portugal, 2007; arriving late, they witnessed the computer are now planning and building. Torun, Poland, 2007; and Keszthaly, Hungary, 2008.

52 The Messenger 135 – March 2009 Astronomical News

Report on the ESO Workshop Large Programmes held at ESO Headquarters, Garching, Germany, 13–15 October 2008

Gautier Mathys1 The first two days were devoted to the enough observing time to approach a Bruno Leibundgut1 presentation of 20 LPs with topics ranging major astrophysical problem over four from the distant Universe and the deter- semesters was generally appreciated. 1 ESO mination of cosmological parameters to the characterisation of the population The bibliometric analysis of LPs com- of nearby galaxies and the search for pared to other programmes shows A report is presented of the workshop habitable exoplanets. Most fields in astro- a good publication record. While 15% of on the progress of ESO Large Pro- physics were represented by an LP. The the observing time was devoted to LPs grammes completed between the last morning of the third day was dedicated they returned 18% of the refereed publi- workshop in May 2003 and September to a discussion of the special scheduling cations. This may be partly due to the 2007. constraints and challenges presented fact that LPs do receive a high priority for by LPs, a bibliometric assessment of the observations. Papers based on LPs scientific impact of LPs and a presenta- appear to have a slightly higher impact — Five years after the first workshop on ESO tion on how to submit the reduced data as measured by the number of citations Large Programmes (see the summary products to the ESO archive. After a bril- per paper — than all other types of pro- by Wagner & Leibundgut, 2004), about liant summary by Willy Benz on the scien- grammes (Normal, Target of Opportunity, 50 participants, including the PIs of the tific value of LPs and how they can suc- Guaranteed Time Observations and second round of Large Programmes ceed (or fail), a discussion, led by the Director Discretionary Time). The effect (LPs), as well as several members of the STC chair, Linda Tacconi, on the various is significant, but not dramatic. It is Observing Programmes Committee aspects of LPs took place. The workshop remarkable that LPs have requested time (OPC) and the Science and Technical programme and the presentations can be on all instruments, with ISAAC, FORS2, Committee (STC) together with some found online1. VIMOS and WFI being the most used. Council members, gathered in Garching The distribution over the different scien- from 13–15 October 2008. The VLT has A number of the projects presented actu- tific categories further reveals a predomi- been in operation for nearly ten years ally encompassed more than one LP, nance of cosmological projects (OPC and a large fraction (15%) of the observ- among them the public surveys (the ESO category A), requesting and being allo- ing time has been devoted to the execu- Imaging Survey [EIS] and the Great cated about half the time. The other three tion of Large Programmes. At the request Observatory Origins Deep Survey categories share the remaining fraction of the OPC, ESO organised this second [GOODS]). There were projects requiring equally. With the extension of LPs on workshop to obtain a new overview of large data samples or deep searches for La Silla to four years (compared to the the scientific results achieved through the extremely rare objects and, in some two years so far on all ESO telescopes) a Large Programmes conducted at the cases, a long time span was essential for marked increase in the time requests for La Silla Paranal Observatory. The work- the observations to record a light curve or HARPS for Period 83 can be noted. The shop featured scientific presentations to measure proper motions or radial veloc- demand for LPs remains high and the of all LPs that were completed between ities. Several LPs complemented other OPC has seen a significant increase in the May 2003 LP workshop and end of large efforts by space- or ground-based the number of LP proposals in the past ESO Period 79 (30 September 2007). The consortia. Very few LPs could be consid- two semesters. teams of investigators leading these LPs ered failures: most of these cases had to were invited to present their scientific cope with instrumental problems, with According to the summary by Willy Benz, results and the impact of their project on the result that the final data quality was LPs need to be bottom-up, i.e., tailored its field. The presentations were followed not sufficiently high to achieve their goals. to the user’s needs. This is guaranteed by a discussion session on the general first through the selection by the OPC, scientific impact of ESO facilities. It was stated several times that LPs have and then by the regular status reports changed some of the culture of astro- provided to the OPC, that enable a judge- One of the outcomes of the May 2003 nomical observations. The need for large ment whether the continued investment workshop was a suggestion that ESO data samples and complex data analyses in telescope resources is still warranted. should archive the legacy data products requires teams with a wide range of Also, LPs should not be regarded in isola- of Large Programmes. This suggestion expertise. This leads to large collabora- tion, but should be seen as a complement was implemented with the requirement tions. Consequently, most LPs were to other types of observing programmes. that Large Programmes that started after granted to large collaborations. Neverthe- They should enable projects, which other- 1 April 2005 deliver Advanced Data Prod- less, there are specific experiments, wise would not be possible at a public ucts (ADPs) to the ESO science archive which can be run by a dedicated small observatory. The results from this work- by the time of publication of their results team concentrating on a specific prob- shop as well as the first workshop in 2003 in a refereed journal. The workshop fea- lem. In most cases the scientific returns show that LPs have enabled European tured a presentation of the ADP submis- of the LPs have been very good, some astronomers to compete on a par with sion process and a discussion of its value even spectacular. The chance to obtain some of the private large facilities in the for the ESO scientific community. US. At the same time, it has to be realised

The Messenger 135 – March 2009 53 Astronomical News

that LPs make use of an expensive The workshop overall was very success- complement the current arsenal of pro- resource and that they have to provide ful and clarified the need for, and the gramme types undertaken with ESO additional benefits for the community. competitive edge of, Large Programmes facilities at the upper end. It is planned to This has led to the requirement that at ESO facilities. The increased demand monitor their success in another work- reduced data from LPs, once published, for the 3.6-metre telescope after the shop in a few years time. should be returned to the ESO Archive time limit for LPs was raised to four years so that they can be used by other astron- speaks for itself. There are some very omers, possibly for different purposes. substantial programmes in progress, References The large investment by the community which will keep this telescope busy for Wagner, S. & Leibundgut, B. 2004, The Messenger, into LPs should justify this modest return. years to come. 115, 41 There was a lively discussion on how this return should be achieved and Beyond the Large Programmes, the Pub- whether it would put astronomers using lic Surveys with VISTA and VST will Notes ESO facilities at a disadvantage com- start during this year and next year. These 1 http://www.eso.org/sci/meetings/LP2008/program. pared to users of private observatories. will be truly massive projects, which html

ESO and the International Year of Astronomy 2009 Opening Ceremony

Pedro Russo1 2 Figure 1. Catherine Cesarky, IAU President, Lars Lindberg Christensen1 addressing the audience during the IYA2009 Opening Ceremony Douglas Pierce-Price1

Yours to Discover”. The IYA2009 marks 1 ESO the 400th anniversary of the first astro- 2 International Astronomical Union nomical observation through a telescope

Credit: IAU/Jose Francisco Salgado by Galileo Galilei. Proclaimed by the United Nations (UN) and endorsed by the The ESO contributions to the Interna- International Council for Science, the tional Year of Astronomy 2009 and IYA2009 has already captured the imagi- the Opening Ceremony, held in Paris in nation of countless individuals. The aim January 2009, are summarised. of the IYA2009 is to stimulate worldwide interest, especially among young people, “The International Year of Astronomy in astronomy and science. Events and 2009 is an important step in furthering the activities will promote a greater apprecia- interest of the public in what is arguably tion of the inspirational aspects of astron- the oldest of all sciences: astronomy.” omy that ESO is keen to foster. ESO Director General, Tim de Zeeuw, on the United Nations proclaiming 2009 as The opening ceremony of IYA2009 was the International Year of Astronomy. held in Paris, on 15–16 January, 2009, under the aegis of the UN, UNESCO and Few areas of science touch on as many the IAU. The ceremony itself featured key- topics of interest for the general public interest means that many organisations, note speeches, research findings, an as astronomy. For countless thousands such as ESO, invest in outreach initiatives. exhibition and also social aspects. About of years, eyes have gazed up at the 900 people attended, among them heavens and wondered; it could be The year 2009 has been launched by the eminent scientists, including Nobel Laure- argued that everyone is born an astrono- International Astronomical Union (IAU) and ates, and also around 100 young students mer. Modern astronomy is a highly pro- the United Nations Educational, Scientific from individual countries. The ceremony fessional field, but is not exclusively the and Cultural Organization (UNESCO) as was very well received, and the quality of domain of specialised scientists. The sky the International Year of Astronomy 2009 talks highly praised. The enthusiasm of all is a shared resource and intense public (IYA2009) under the theme “The Universe, involved shone through, and proved that

54 The Messenger 135 – March 2009 Credit: IAU/Lee Pullen Credit: ESO/D. Pierce-Price

Figure 4. The Interna- tional Year of Astronomy 2009 logo.

Figure 3. The ESO exhibition took prominent posi- coordinate a 24-hour webcast from tion in the display room. research observatories; the Cosmic Diary, a blog that gives the public insight into the Paranal in Chile. This was a rare opportu- lives of professional astronomers; and the nity for the audience to join astronomers Portal to the Universe, a global one-stop Figure 2. A video link with Paranal made a big impres- at work in the VLT control room, towards website for online astronomy content. sion on the audience. Here, ESO Director General Tim the end of the Chilean night. Christophe de Zeeuw is seen talking to Christophe Dumas on the big screen, live from the VLT Control Room. Dumas, head of the Science Operations The IYA2009 Secretariat at ESO HQ is the Department at the VLT, discussed the central hub of the IAU’s implementation the IYA2009 has already gained plenty of night’s observations and shared a recent of the IYA2009, and was established to momentum to help popularise astronomy image from the telescope. You can watch coordinate activities during the planning, throughout 2009, and beyond. this on the IYA2009 Opening Ceremony execution and evaluation of IYA2009. The video archive1. IYA2009 science writer, secretariat liaises continuously with the Lee Pullen, kept a “Cosmic Diary Live- — currently 137 — national nodes, task ESO’s contribution to the ceremony Blog” during the ceremony. This diary of groups, partners and organisational asso- events was regularly updated, and fol- ciates, the media and the general public ESO held an exhibition at the UNESCO lowed by people in over 130 countries. to ensure the progress of IYA2009 at all headquarters in Paris as part of the open- levels. The secretariat is staffed by Pedro ing ceremony. The exhibition presented Russo, the IYA2009 coordinator, IYA2009 information about ESO’s role in astronomy ESO’s commitment to public outreach assistant Mariana Barrosa and is man- and the Very Large Telescope (VLT), the and to IYA2009 aged by Lars Lindberg Christensen, head European Extremely Large Telescope of ESO’s education and Public Outreach (E-ELT) and the Atacama Large Millimeter/ Since plans for IYA2009 were first laid in Department. ESO’s own organisational submillimeter Array (ALMA) projects. A 2003, ESO has played a major role in IYA2009 contact is Douglas Pierce-Price. scale model of the E-ELT complemented the project. ESO is hosting the IYA2009 the exhibition. The official IYA2009 bro- Secretariat for the International Astronom- Astronomy is one of the oldest fundamen- chure, an E-ELT brochure, and “Eyes on ical Union, which coordinates the activi- tal sciences, yet continues to make a the Skies” DVDs were available for visitors. ties globally. ESO is one of the Organisa- profound impact on our culture and is a tional Associates of IYA2009, and was powerful expression of the human intel- The afternoon session on Thursday, 15 also closely involved in the resolution sub- lect. 2009 is the year in which individuals January, was chaired by the ESO Director mitted to the UN by Italy, which led to the and organisations can make a difference General Tim de Zeeuw. The session saw 62nd General Assembly of the UN pro- by popularising astronomy as never before two fascinating talks: “The New Frontier: claiming 2009 as the International Year of and really “bring the Universe down to The Exploration of the Solar System” Astronomy. Earth”. ESO’s presence at the opening by André Brahic; and “Echoes of Creation: ceremony shows our commitment to this Discovery of the Big Bang Fossil Radia- In addition to a wide array of activities grand aim. tion”, by Nobel Laureate Robert Wilson. planned both at the local and international A highlight on Friday, 16 January, was Tim level, ESO is leading three of the eleven de Zeeuw’s overview of ESO activities global Cornerstone projects. These are Notes and the live video conference from the 100 Hours of Astronomy, a worldwide 1 http://www.astronomy2009.org/webcast auditorium in Paris to ESO’s VLT on Cerro observing marathon during which ESO will

The Messenger 135 – March 2009 55 Astronomical News

Announcement of the ESO Workshop MAD and Beyond: Science with Multi-Conjugate Adaptive Optics Instruments

8–10 June 2009, ESO Headquarters, Garching, Germany

ESO has pioneered the use of adaptive The study phase for the European have been published and many others optics assisted instruments for research Extremely Large Telescope (E-ELT) pro- are in preparation, showing that the in astronomy. Come-On+ and its heir, vided the opportunity to actually demon- community is keen to apply MCAO Adonis, were the first common-user adap- strate that atmospheric tomography, as techniques to ambitious astronomical tive optics instruments. Nowadays adap- implemented in its best known version of problems. tive optics (AO) instruments are standard Multi-Conjugate Adaptive Optics (MCAO), at all major observatories. AO adapters can provide AO performance over fields The success of MAD also demonstrated are routinely used to feed spectrographs of view significantly larger than the isopla- that the technology is mature, and the that require very small entrance windows natic patch. This led to the construction community prepared for the next genera- to achieve very high spectral resolution, of the Multi-conjugate Adaptive-optics tion of MCAO instruments. The aims of such as CRIRES, or to increase the spa- Demonstrator, MAD, for the VLT. Thus, the workshop, therefore, will be twofold: tial resolution of spectro-imagers, such about 20 years after the deployment of to celebrate the achievements of MAD as SINFONI on the Very Large Telescope the AO demonstrator Come-On at the through a dedicated discussion of its (VLT). In addition AO adapters are indis- 3.6-metre telescope at La Silla, ESO is design constraints and scientific achieve- pensable for the Very Large Telescope again pioneering the field by commission- ments, and, with the strong foundation Interferometer (VLTI). The biggest short- ing MAD on Unit Telescope 3 at the VLT. provided by the scientific results, to outline coming of AO instruments is their small The commissioning was so successful, the high level requirements for the next corrected field of view, which is limited that strong demand from the community generation of MCAO instruments. Thus, by the size of the isoplanatic patch that led MAD to be offered for scientific the spirit of the conference will be both even in the infrared and at the best sites observations, initially for 14 nights in the festive, because we are celebrating MAD, rarely exceeds 15 arcseconds. The ESO Chilean summer of 2007/2008 and then, and visionary, because we are dreaming Workshop on AO in Venice in 2001 paved again at the request of the community, about the future. the way, on the basis of theory and simu- for an additional nine nights in August lations by researchers in Europe and the 2008 to cover the winter period. One year Further details can be found at www.eso. USA, to overcoming the isoplanatic barrier after the first science demonstration run, org/sci/meetings/mad2009/index.html. and thus atmospheric tomography was close to ten papers based on MAD data The deadline for registration is 10 May born. 2009.

Announcement of the IAU Special Session 1 IR and Sub-mm Spectroscopy: A New Tool for Studying Stellar Evolution

3–6 August 2009, IAU General Assembly, Rio de Janeiro, Brazil

and observational astrophysics, instru- early results (Herschel) or making ad­­ mentation and laboratory spectroscopy. vanced preparations for launch (SOFIA, In combination, these fields hold the key JWST). New ground-based facilities for the scientific success of current and (VLT, Gemini, Keck) have matured to the planned facilities. New observations will point of presenting results of unprece- foster new approaches to old problems dented quality. In the near future ALMA and will no doubt lead to transformational will open up the study of sub-millimetre thinking on stellar evolution. sources to unprecedented sensitivity. The next generation of extremely large The conference occurs at a particularly telescopes (ELTs) will allow the study of advantageous time for the transfer of individual stars in other galaxies. In this knowledge in IR and sub-millimetre spec- field, IR spectros­copy will be particularly troscopy from mission to mission. Certain important, as only the combination of space missions have produced a wealth ELTs with active optics will allow indi­vidual of data — Spitzer will have completed stars to be singled out in crowded regions. its cryogenic mission and entered the This IAU Special Session aims at fostering “warm” phase, while AKARI has ended For details please visit the website at: collaboration between various fields and operations after completing an all-sky www.eso.org/sci/meetings/iau2009-sps1/ will bring together experts from theoretical survey. Others will either be presenting index.html

56 The Messenger 135 – March 2009 Astronomical News

Announcement of the ESO Workshop Detectors for Astronomy 2009

12–16 October 2009, ESO Headquarters, Garching, Germany

Astronomical observations are critically Figure 1. The 8 x 4 dependent on focal-plane array technol- mosaic of e2v CCD82-44 2K x 4K ogy, and detectors continue to play a key CCDs (16K x 16K pixels role in continuing to extend the scope in total) for OmegaCAM, of astronomical observations. Higher sen- which will capture the sitivity, reduced noise, larger formats, full 1 degree x 1 degree field of view of the VLT better cosmetic quality, higher quantum Survey Telescope (VST). efficiency, smaller point-spread functions, lower dark current, higher bandwidth, and many more, constantly set new mile­- stones on the roadmap towards the goal of artefact-free photon shot noise limited images of reality. One of the fastest growing applications is signal sensing, especially wavefront sensing for adaptive optics and fringe tracking for interfer­ ometry, which have become vital enabling The 2009 Workshop Detectors for or novel requirements on detectors technologies for both interferometry and Astronomy aims at providing an up-to- – Scientific applications and results that extremely large telescopes. Topics of date platform for such exchanges and depend on high performance detectors active research are large format Comple- continues a series of similar meetings – Test methodology and quality control mentary Metal Oxide Semiconductor in 1991, 1993, 1996, 1999 (all at ESO- – Calibration of performance (CMOS) and Charge Coupled Devices Garching), 2002 (Waimea), and 2005 (CDD) array mosaics, orthogonal transfer (Taormina). The 2009 meeting will spe­ Contributions are invited irrespective of CCDs, electron multiplication CCDs, elec- cifically address the following topics: wavelength and deployment on the tron avalanche photodiode arrays, quan- ground or in space. The main focus will tum-well infrared photon detectors, – Detector technologies and design be on the optical and infrared domains. Application Specific Integrated Circuits – Detector manufacturing Depending on interest, splinter meetings (ASICs), blocked-impurity band arrays, – Detector evaluation and calibration dealing with topics of special interest novel readout technologies, to name a few. – Control electronics can be organised. Contributions with In a field with such rapid and complex – ASICs demonstrations of hard- or software are developments, it is essential that design- – Control software welcome (subject to technical feasibility). ers, manufacturers and users gather – Detector systems regularly in order to exchange information – Mosaic focal-plane arrays For registration and more information about requirements, technical possibilities – Cryo-vacuum technologies please visit www.eso.org/sci/meetings/ and achievements on a worldwide scale. – Instruments with very demanding and/ dfa2009/.

ESO’s Studentship Programmes: Training Tomorrow’s Astronomers Today

Michael West1 facing declining student enrollments, ESO is a leader in shaping the future Marina Rejkuba1 the ASTRONET strategic plan for of astronomy, and one important way Bruno Leibundgut1 European Astronomy notes “young stu- to achieve this goal is by offering short- Eric Emsellem1 dents have continued to enter the term and long-term studentships that field at a steady level”. Indeed, with provide excellent opportunities for stu- Very Large Telescope (VLT), Atacama dents to pursue research under the 1 ESO Large Millimeter/submillimeter Array supervision of ESO staff astronomers. (ALMA) the European Extremely Large Since its inception two decades ago, Telescope (E-ELT) and other exciting hundreds of young astronomers have Students are the lifeblood of astronomy, new facilities on the horizon, it is spent some time during their PhD pro- the next generation of astronomers. hard to imagine a better time to be an gramme at ESO in Garching or Santiago. While other scientific disciplines are astronomy student. Many have gone on to leading positions

The Messenger 135 – March 2009 57 Astronomical News

Figure 1. (Left) Group portrait of students at ESO Garching.

Figure 2. (Right) Students at ESO Chile enjoy an excursion to Paranal.

by the Director General Discretionary Fund, which supports a broad range of needs related to scientific projects by ESO staff astronomers.

There is no formal application procedure for short-term studentships at ESO. Students interested in the possibility of working on a short-term research project at ESO should explore the science staff web pages to identify one or more ESO staff astronomers with whom they might wish to work and then contact those astronomers directly, or contact the Head of the Office for Science in Garching or Santiago for more information. at universities, observatories and other organisations in Europe and beyond. Students in Santiago Students in Garching PhD Studentships Surrounded by the splendour of the Currently 22 students from 15 countries Andes Mountains, seven students are Through its PhD studentship programme, are working on their PhD research at currently pursuing their PhD research at ESO offers training to future users of ESO Garching (see Figure 1). Fifteen of ESO offices in Santiago, together with its state-of-the-art observational facilities. them are enrolled in the ESO studentship a constant stream of short-term students Students that come to ESO through the programme and will spend one or two (Figure 2). During the past year alone, stu- PhD studentship programme are already years at ESO during their PhD and will dents from universities in Belgium, Chile, enrolled in doctoral programmes in receive the degree from their home uni- Denmark, England, France, Germany, astronomy, physics or related fields at versity. One student is participating in an Iceland, Italy, Portugal and the United universities in ESO member states or non- exchange programme with Chinese uni- States have participated in ESO Chile’s member state institutes. versities, and six students are enrolled in short-term and long-term studentship the International Max-Planck Research programmes. ESO offices in Chile provide ESO PhD students typically spend one School (IMPRS) on Astrophysics. a stimulating scientific environment for or two years pursuing their doctoral students, with a science team of nearly 80 research under the co-supervision of an ESO’s Garching Headquarters forms staff astronomers, fellows, students ESO faculty astronomer, in close contact part of one of the world’s largest centres and visitors, plus close connections with with activities and people at one of the for astronomy and physics, with other ALMA and the astronomy departments world’s foremost observatories. After leading institutes such as the Max-Planck at top Chilean universities in Santiago and returning to their home universities, these Institute for Astrophysics, the Max-Planck beyond. students are often expert users of ESO Institute for Extraterrestrial physics, observing facilities. Upon the completion MPE, the Institute for Plasma Physics, the In addition to engaging in scientific of their PhD a large majority of these stu- Max-Planck Institute for Quantum Optics re­search under the supervision of an dents pursue a career as a professional and the campus of the Munich Technical ESO staff astronomer, students at ESO astronomer, and some have become University nearby. The strength of this Chile have a unique opportunity to get postdoctoral fellows or staff astronomers concentration of scientific expertise was hands-on experience with the technical at ESO. recently recognised by the German aspects of ob­­servatory operations at authorities, and led to the creation of the Paranal, La Silla, ALMA or the Atacama Excellence Cluster on the “Origin and Pathfinder Experiment (APEX), a millime- Short-term studentships Structure of the Universe”. In such a tre/submillimetre telescope. This month, vibrant environment students have an for example, ESO Chile PhD student Opportunities also exist for a limited abundance of choice for seminars and Pedro Almeida will travel to La Silla to number of students to come to ESO for lectures with unique opportunities to spend two weeks working on polarimetry periods of one to three months to work learn about the hottest topics and most data from EFOSC2 and participating in on research projects with ESO astrono- important open questions in physics and observations using HARPS, all under the mers. Such positions are usually funded astronomy. guidance of ESO astronomers.

58 The Messenger 135 – March 2009 Astronomical News

ESO

European Organisation for Astronomical Research in the Southern Hemisphere

ESO Studentship Programme

The European Southern Observatory research student programme The closing date for applications is 15 June 2009. aims to provide opportunities to enhance the PhD programmes of Late applications will be accepted until all the positions are filled. ESO member-state universities. Its goal is to bring young scientists into close contact with the activities and people at one of the world’s Please attach to your application the following documents: foremost observatories. For more information about ESO’s astro­ nomical research activities please consult www.eso.org/science/. – a Curriculum Vitae (including a list of publications, if any), with a copy of the transcript of university certificate(s)/diploma(s); The ESO studentship programme is shared between the ESO Head- quarters in Garching (Germany) and the ESO offices in Santiago – a summary of the Masters thesis project (if applicable) and ongo- (Chile). These positions are open to students enrolled in a PhD pro- ing projects, indicating the title and the supervisor (maximum half gramme in astronomy or related fields. In addition, ESO will provide a page), as well as an outline of the PhD project, highlighting the up to two studentship positions per year in Santiago for students advantages of coming to ESO (recommended 1 page, max. 2); enrolled in South American universities. – two letters of reference, one from the supervisor/advisor at the Students in the programme work on their doctoral project under home institute and one from the ESO local supervisor; the formal supervision of their home university. They come to either Garching or Santiago for a stay of normally between one and two – a letter from the home institution that: i) guarantees financial sup- years to conduct part of their studies under the co-supervision of an port for the remaining PhD period after the completion of the ESO ESO staff astronomer. Candidates and their supervisors at the home studentship; and ii) indicates whether the requirements to obtain institute should agree on a research project together with the ESO the PhD degree at the home institute have already been fulfilled. local supervisor. A list of potential ESO supervisors and their research interests can be found at www.eso.org/sci/activities/personnel.html. All documents should be submitted in English (but no translation is A list of current PhD projects offered by ESO staff is available at required for the certificates and diplomas). www.eso.org/sci/activities/thesis-topics/. It is highly recommended that the applicants start their PhD studies at their home institute Review of the received material, including the recommendation let- before continuing their PhD work and developing observational ters, will start on June 15. Applications arriving after this deadline expertise at ESO. will be considered until all the positions are filled. Incomplete appli- cations will not be considered. All reference letters should be sent ESO Chile students will have an opportunity to visit the observato- electronically to [email protected]. ries and get involved in small projects aimed at providing an insight into operations at the observatory. Such visits and projects, which Candidates will be notified of the results of the selection process in are voluntary, will be scheduled to ensure that they do not interfere July 2009. Studentships typically begin between August and with the research project of the student in Santiago. December of the year in which they are awarded. In well-justified cases, starting dates in the year following the application can be In Garching, students can benefit from the series of lectures given to negotiated. the PhD students enrolled in the IMPRS (International Max-Planck Research School on Astrophysics) PhD programme; attendance is For further information please contact Christina Stoffer (cstoffer@ voluntary. Students who are already enrolled in a PhD programme in eso.org). the Munich area (e.g., the IMPRS or at a Munich University) and wish to apply for an ESO studentship in Garching, should provide com- The post is open equally to suitably qualified male and female appli- pelling justification for their application. cants.

The Outline of the Terms of Service for Students (www.eso.org/ public/employment/student.html) provides some more details on employment conditions and benefits.

The Messenger 135 – March 2009 59 Astronomical News

New Staff at ESO

Jonathan Smoker

Although born in Chicago, I was brought up on a smallholding in rural England. Amongst the chickens, sheep, goats, windmill and methane digester could be found a 6-inch Newtonian telescope that my Dad had made out of a plastic tube and a mirror that someone had given us. So, that’s how I became an astronomer! Although I can’t quite remember the Moon landings, one thing that I do recall clearly came later, in the 1970s, when Viking touched down on Mars and “dis- covered” life there, and later on the first shuttle launch.

I went to university in Manchester for my Jonathan Smoker undergraduate degree, then did my PhD in radio astronomy at Jodrell Bank, study- ing the neutral hydrogen content of low Although working on Paranal can be tir- for quite some years, I am enjoying the surface brightness and blue compact gal­ ing, it is balanced by the satisfaction new challenges presented by the Atacama axies, but also doing a bit of imaging of being part of a team that has contrib- Large Millimeter/submillimeter Array and spectroscopy. After a six-month spell uted so much to astronomy over the last (ALMA) project and this large international living in a tent at Jodrell, I switched to 10-plus years. There is nearly always collaboration. system administration at the Institute of something new happening on the moun- Astronomy in Cambridge, before moving tain (APE is testing on Melipal and a Working and living in the Munich area to Queen’s University, Belfast, working GRB Target of Opportunity has arrived again is a pleasure after 18 years of “ven- on the Magellanic Bridge and high velocity as I write these words)... may that stay turing” through the world at different clouds. I went on 22 observing trips in the same for years to come! institutes and observatories. After study- four years, giving me enough experience ing physics in Munich and obtaining a to be able to apply for a staff position PhD on submillimetre instrumentation and at ESO in 2002. After a year or so I was Wolfgang Wild astronomy in 1990, I moved to Chile to lucky enough to be involved in the science work at SEST (Swedish–ESO Submillime- verification of FLAMES, for which I later In November 2008 I took up the position tre Telescope) as an ESO fellow on La became instrument scientist. It still of Head of the ALMA Division and Euro- Silla. These were interesting times both at amazes me how the robot can place so pean ALProject Manager. Having worked the telescope and in Chile — the country many fibres with an accuracy of a fraction in the field of sub-/millimetre and far- had just made the transition to a democ- of an arcsecond! After a spell back in infrared instrumentation and astronomy racy. I much enjoyed the variety of tasks Belfast and travelling in South America whilst working on my laptop, I returned to ESO in 2008, again initially working on FLAMES-UVES, but who knows what the future will hold? One of the great things about ESO is that there are always oppor- tunities to work on new projects, although balancing observatory work, science and the odd trip to the Andes is always a challenge!

My research is concerned with probing the tiny-scale (astronomical unit) structure of the interstellar medium, using UVES to observe early type stars at high resolu- tion and twin epochs. I have also dabbled as a collaborator in high latitude stars and typing the precursors of supernovae.

Wolfgang Wild

60 The Messenger 135 – March 2009 in the small SEST team, ranging from was still called “MMA–LSA” (Millimeter these early ALMA days, from 1999 to helping visiting astronomers and doing Array/Large Southern Array), which was 2002, I was also the European Receiver my own astronomical research to im­­ a merger between two similar, but differ- Team Leader working closely with col- proving the system and solving technical ent, interferometer projects in the US and leagues from North America and Japan problems (I remember once crawling Europe. Following various meetings and on the design of the ALMA receivers. It inside the antenna structure at night trying discussions, the MMA–LSA evolved into looked like I would stay with ALMA for to find a short circuit). The good working the ALMA project, and with the participa- quite a while, but then in 2002 there was atmosphere at La SiIla and the friendly tion of East Asia (Japan and Taiwan) and the opportunity to contribute to a space people and natural beauty of Chile also Canada, ALMA developed into what it is instrument in the same institute, and I contributed to memorable years. today — a large international collaboration worked on HIFI (Heterodyne Instrument to build a new interferometer at a high site for the Far-Infrared), one of the three When the ESO fellowship came to an end in Chile. Wanting to contribute to ALMA instruments onboard ESA’s Herschel I had the opportunity to become Site Man- on the instrumentation side, I heard of a Space Observatory, foreseen for launch ager of the Pico Veleta Observatory near new group being built up in the Nether- on an Ariane 5 rocket in April this year. Granada in southern Spain. The observa- lands for the development of the ALMA In 2004 I was appointed Head of SRON’s tory belongs to the French–German– Band 9 receivers, one of the challenging Low Energy Astrophysics Division (with Spanish Institute IRAM (Institut de Radio- high frequency bands. I applied for the HIFI being the major project), and from astronomie Millimetrique) with its head- position of Project Manager and group 2007 onward I became responsible for quarters in Grenoble (France). During my leader and was hired as the first person in SRON’s infrared and submillimetre pro- years at the IRAM 30-metre telescope, this new undertaking between NOVA, the gramme as Programme Scientist. During I was responsible for the operations and SRON Netherlands Institute for Space my time at SRON I also held a position improvements of the observatory together Research and the University of Groningen. at the Kapteyn Astronomical Institute of with a group of scientific and technical the University of Groningen, and in 2005 experts — an interesting job that taught Those early ALMA days were quite excit- I was appointed Professor for Techniques me a lot. I also enjoyed life in Granada and ing — I started by hiring another person, of Far-Infrared and (Sub-)Millimetre married a “Granadina”. so we began with two people, many Astronomy. ideas, an empty lab, a budget (never However, after nine years of working at enough of course) and a lot of enthusi- I am very happy to be part of ESO again, telescopes, by around 1999 it was time for asm. Today, almost ten years later, the and with such exciting projects like ALMA a change and I wanted to go back to group in Groningen comprises twelve and the European Extremely Large Tele- instrumentation development. At that time, people and has built nine state-of-the-art scope I am looking forward to these great ALMA was just about to become a joint receivers for ALMA with another 60+ to new observatories and the science they project between Europe and the US. It be built over the next few years. During will enable. Credit: ESO/L. Testi

Sunset over the ALMA Test Facility (ATF) in Socorro, New Mexico. The prototype US and European antennas, shown here, were used for testing and development of hardware and software and were interferometrically linked (see ESO PR 10/07). The ATF closed in December 2008, having suc- cessfully fulfilled its purpose.

The Messenger 135 – March 2009 61 Astronomical News

Fellows at ESO

managing technical challenges, we be trained on two UTs, Antu and Kueyen, have succeeded in getting unprecedented performing over 1500 OBs for a cumula- astrophysical results. I am now in the tive open shutter time of 587 hours. Given fourth year of my ESO fellowship and the ratio of the instrument aperture sizes, thinking about future projects and the next it would have taken 17 290 years for Gali- generation of VLTI instrumentation that leo to collect as many photons as I did. will use the four giant telescopes together. But the most impressive part is not about My wife and I have enjoyed Chile numbers, sizes or advanced technology immensely — the people, nature, food, (although they are very impressive), it and all the new friends we now have. is about the people, working around the Moreover, we will leave Chile with two clock, away from family, often tired everlasting souvenirs: our son Tobie and because of the workload, stress and lack our daughter Anaelle. of sleep. It is about their problem-solving attitude, their constant mood and their incredible motivation. My ESO fellowship Hugues Sana has certainly been a wonderful challenge Jean-Baptiste Le Bouquin and one of the most intense life experi- 11961 km... That’s the distance between ences. If I had to summarise what I Liège and Santiago. 11961 km... it is diffi- learned in a single word, it would be “bal- Jean-Baptiste Le Bouquin cult to get further away. It is further than ance”. Balance between observatory the US West coast, Hong Kong or Beijing, duties, science time, private life and per- I started astronomy in quite an unusual and even further than Tokyo. The only sonal development, and what it takes, way. I first studied heat engines and other way to go further would be to head for as a juggler, to keep that many balls flying sorts of machinery using thermal energy. Australia, to some small Pacific island or in the air. Having been lured into an increasing to Antarctica. passion for physics, I continued studying About me — I obtained my PhD at Liège fundamental physics for several more 11961 km... it is not the other side of the University in 2005, working on optical years. world, but it is getting close and that is and X-ray spectroscopy of massive stars. the adventure I had the chance to take up After an additional year in Liège, I moved There came a time when I had to choose about three years ago when I was offered to ESO/Chile in May 2006. In Paranal, between particle physics and astrophys- a fellowship position at ESO. It was on I first joined the CRIRES instrument team ics. My wife found the best strategy a Wednesday morning, two days before before taking over the enjoyable task to make this dramatic choice: look at the Christmas, 1145 days ago and 11961 km of UVES instrument scientist, where I am potential colleagues in both areas and away from where I am sitting right now. learning what it takes to have an instru- assess how they were set up in life (prop- ment working to the best of its perform- erly dressed or not, with family or not, Since then, I have enjoyed 54 Lan Chile ance. In Santiago, I pursue my own etc.). The “vote” definitely went to astro- snack boxes, delighted in 233 wonderful research projects on massive stars, tak- physics! sunsets in Paranal and survived 671 rides ing advantage of the active collaboration from my house in Santiago to the Vitacura of a dozen colleagues working in the I completed my PhD thesis in the city of office. In Paranal, I had the opportunity to nearby field of stellar clusters. Grenoble, in the French Alps, conveniently located near numerous ski resorts. I visited Paranal once during my graduate studies, where I worked on the VLTI and was definitely impressed by this huge machine. Press a button and you have two giant telescopes pointing together and unveiling the mysteries of some unknown stars. I decided to be part of this adventure.

I arrived at ESO in April 2006. I devoted a significant part of my duties and research time to improving the abilities of the VLTI. I greatly appreciated sharing the work with engineers, technicians and astronomers. I am proud of what has been achieved within these last three years. As well as Hugues Sana

62 The Messenger 135 – March 2009 Personnel Movements

Arrivals (1 January–31 March 2009) Departures (1 January–31 March 2009) Europe Europe Zinsmeyer, William (US) Software Engineer Puech, Florence (FR) VLTI System Engineer Micol, Alberto (IT) VO and Astronomical Data Specialist da Rocha, Cristiano (BR) Fellow Chasiotis, Stella-Maria (DE) Secretary Mazzoleni, Ruben (IT) Paid Associate Vraux, Jean Pierre (FR) Contract Office Lorch, Henning (DE) Software Engineer Verzichelli, Gianluca (IT) System Engineer Stöckl, Josef (AT) Student Baginksi, Isabelle (DE) Administrative Assistant Lopez Silva, Joao (CL) Student Fourie, Petrus Gerhardus (ZA) System Engineer Sharkey, Colleen (US) Hubble Outreach Coordinator Hekman, Peter (NL) Electronic Engineer Paolo Ghiretti (IT) Civil Engineer Emsellem, Eric (FR) Head of the Office for Science - Garching Ferguson, Neil (GB) Software Engineer Singh, Paul (IT) Junior Optical Engineer Argomedo, Javier (CL) Software Engineer Ruiz Velasco, Alma (MX) Student Guidetti, Daria (IT) Student Lopez Silva, Joao (CL) Student Hoffstadt Urrutia, Arturo (CL) Student McNeil, Emily Kate (US) Student Mueller, Eric (DE) Student Daemgen, Sebastian (DE) Student Zheng, Zheng (CN) Student Müller, André (DE) Student

Chile Chile Kneissl, Ruediger (DE) Operations Astronomer Hubrig, Swetlana (DE) Operations Astronomer Vila Vilaro, Baltasar (ES) Systems Astronomer Mirabel, Igor-Feliz (FR) Representative of ESO - Chile Poupar, Sébastien (FR) Mechanical Engineer Gallardo, Javier (CL) Software Engineer Rawlings, Mark (GB) Operations Astronomer Sanzana, Lilian Software Engineer Rengaswamy, Sridharan (IN) VLTI Astronomer Naef, Dominique (CH) Fellow Evatt, Matthew (US) Mechanical Engineer Sahlman, Johannes (DE) Student Oestreich, Martin (DE) Electrical Engineer Ramírez, Christian (CL) Mechanical Engineer Díaz, Alvaro (CL) Instrument Technician Montagnier, Guillaume (FR) Fellow Alvarez Candal, Alvaro (AR) Fellow Asmus, Daniel (DE) Student Credit: ESO/FOTOAG

The Garching science campus, home to the Munich Technical University, four Max Planck Institutes, computing and research institutes, and ESO, shown in an aerial photograph taken in July 2007. The tracks of the U-Bahn (underground city railway) are seen to the bottom left (west) and the River Isar to the upper right (east). The dome just above image centre is the old research reactor, with the new reactor housed in the large grey building to its right. The ESO Headquarters building is towards the lower left of the photograph with the storage hall to the east. The ESO extension building will occupy parts of the field to the south (see Figure 3 of Fischer & Walsh, p. 4).

The Messenger 135 – March 2009 63 Annual Index 2008 (Nos. 131–134)

Subject Index E-ELT and the Cosmic Expansion History – A Far Behind the Scenes of the Discovery of Two Stretch?; D’Odorico, V.; Levshakov, S.; Bonifacio, Extrasolar Planets: ESO Large Programme 666; P.; Bouchy, F.; Wiklind, T.; Queloz, D.; Udry, S.; Wyrzykowski, L.; Soszynski, I.; Szewczyk, O.; The Organisation Moscardini, L.; Molaro, P.; Murphy, M.; Lovis, C.; Szymanski, M.; Ulaczyk, K.; Kubiak, M.; Shporer, D’Odorico, S.; Zucker, S.; Mayor, M.; Pepe, F.; A.; Tamuz, O.; Diaz, R.; Zoccali, M.; Ruiz, M. T.; The Perfect Machine; de Zeeuw, T.; 132, 2 Cristiani, S.; Viel, M.; Haehnelt, M.; Pasquini, L.; Gieren, W.; Pietrzynski, G.; Ramirez, S.; Hoyer, S.; 10th Anniversary of First Light of the VLT; 132, 4 Dessauges, M.; Vanzella, E.; Liske, J.; Grazian, A.; Udry, S.; Mayor, M.; Mazeh, T.; Gillon, M.; Queloz, 133, 10 D.; Santos, N.; Moutou, C.; Naef, D.; Udalski, A.; Austria Declares Intent to Join ESO; 132, 5 The Quest for Near-infrared Calibration Sources for Minniti, D.; Pont, F.; Melo, C.; 133, 21 The ASTRONET Infrastructure Roadmap: A Twenty E-ELT Instruments; J. Sansonetti, C.; Ralchenko, Probing Sagittarius A* and its Environment at the Year Strategy for European Astronomy; Bode, M.; Y.; Kerber, F.; Nave, G.; Bristow, P.; Aldenius, M.; Galactic Centre: VLT and APEX Working in Monnet, G.; ASTRONET Roadmap Working 133, 14 Synergy; Wiesemeyer, H.; Zamaninasab, M.; Group; 134, 2 Detector Upgrade for FLAMES: GIRAFFE Gets Red Zensus, A.; Vogel, S.; Thum, C.; Straubmeier, C.; Eyes; Gieles, M.; Palsa, R.; Bendek, E.; Peña, E.; Sjouwerman, L.; Schuster, K.; Moultaka, J.; Mužić, K.; Najarro, F.; Pott, J.; Meyer, L.; Mauerhan, J.; Telescopes and Instrumentation Castillo, R.; Hanuschik, R.; Naef, D.; Deiries, S.; Melo, C.; Pasquini, L.; Downing, M.; 133, 17 Markoff, S.; Lu, R.; Kunneriath, D.; Krips, M.; König, S.; Krichbaum, T.; Karas, V.; Duschl, W.; Advanced Calibration Techniques for Astronomical The VLTI PRIMA Facility; van Belle, G. T.; Sahlmann, Downes, D.; Dovčiak, M.; Morris, M. R.; Weiss, A.; Spectrographs; Bristow, P.; Kerber, F.; Rosa, M. J.; Abuter, R.; Accardo, M.; Andolfato, L.; Brillant, Baganoff, F.; Witzel, G.; Bertram, T.; García-Marín, R.; 131, 2 S.; de Jong, J.; Derie, F.; Delplancke, F.; Duc, T. M.; Schödel, R.; Eckart, A.; 133, 26 Laser Guide Star Adaptive Optics without Tip-tilt; P.; Dupuy, C.; Gilli, B.; Gitton, P.; Haguenauer, P.; Stellar Populations of Bulges of Disc Galaxies in Davies, R.; Rabien, S.; Lidman, C.; Le Louarn, M.; Jocou, L.; Jost, A.; Di Lieto, N.; Frahm, R.; Clusters; Bertola, F.; Sarzi, M.; Pizzella, A.; Kasper, M.; Schreiber, N. M. F.; Roccatagliata, V.; Ménardi, S.; Morel, S.; Moresmau, J.; Palsa, R.; Méndez-Abreu, J.; Maria Corsini, E.; Coccato, L.; Ageorges, N.; Amico, P.; Dumas, C.; Mannucci, F.; Popovic, D.; Pozna, E.; Puech, F.; Lévêque, S.; Morelli, L.; Saglia, R.; Pompei, E.; 133, 31 131, 7 Ramirez, A.; Schuhler, N.; Somboli, F.; Wehner, S.; ESPRI Consortium, T.; 134, 6 Mid-infrared Interferometry of Active Galactic Nuclei: DAZLE on the VLT; McMahon, R.; Parry, I.; an Outstanding Scientific Success of the VLTI; Venemans, B.; King, D.; Ryan-Weber, E.; Bland- News on the Commissioning of X-shooter; Jaffe, W.; Röttgering, H.; Burtscher, L.; Tristram, Hawthorn, J.; Horton, A.; 131, 11 D’Odorico, S.; 134, 12 K.; Schartmann, M.; Raban, D.; Meisenheimer, K.; Report on the JENAM 2008 Meeting Symposium Phase Correction for ALMA: Adaptive Optics in the 133, 36 Submillimetre; Nikolic, B.; Richer, J.; Hills, R.; Stir- Science with the E-ELT; Monnet, G.; 134, 14 The Supernova Legacy Survey; Balland, C.; Sullivan, ling, A.; 131, 14 M.; 133, 42 Hawk-I – First Results from Science Verification; Astronomical Science From the Dynamics of Cepheids to the Milky Way Doherty, M.; Wehner, E.; Willis, J.; Seifahrt, A.; Rotation and the Calibration of the Distance Preibisch, T.; McCaughrean, M.; Mora Fernandes, A Multi-Wavelength Study of the 2003–2006 Scale; Nardetto, N.; Kervella, P.; Barnes, T.; A.; Gieles, M.; Norris, M.; Larsen, S.; Kuntschner, Outburst of V1647 Orionis; van den Ancker, M.; Bersier, D.; Fokin, A.; Fouqué, P.; Gillet, D.; Groh, H.; Kneib, J.; Venemans, B.; Lidman, C.; Kissler- Fedele, D.; Petr-Gotzens, M.; Rafanelli, P.; 131, 20 J.; Kraus, S.; Mathias, P.; Mérand, A.; Millour, F.; Patig, M.; Fontana, A.; 132, 7 The VLT-FLAMES Survey of Massive Stars; Evans, Mourard, D.; Stoekl, A.; 134, 20 Seeing is Believing: New Facts about the Evolution C.; Hunter, I.; Smartt, S.; Lennon, D.; de Koter, STRESS Counting Supernovae; Botticella, M. T.; of Seeing on Paranal; Lombardi, G.; Sarazin, M.; A.; Mokiem, R.; Trundle, C.; Dufton, P.; Ryans, R.; Cappellaro, E.; Riello, M.; Greggio, L.; Benetti, S.; Melnick, J.; Navarrete, J.; 132, 11 Puls, J.; Vink, J.; Herrero, A.; Simón-Díaz, S.; Patat, F.; Turatto, M.; Altavilla, G.; Pastorello, A.; EFOSC2 Episode IV: A New Hope; Snodgrass, C.; Langer, N.; Brott, I.; 131, 25 Valenti, S.; Zampieri, L.; Harutyunyan, A.; Pignata, Saviane, I.; Monaco, L.; Sinclaire, P.; 132, 18 Seeking for the Progenitors of Type Ia Supernovae; G.; Taubenberger, S.; 134, 25 Two Volume-phased Holographic Grisms Now Patat, F.; Chandra, P.; Chevalier, R.; Justham, S.; Swift, VLT and Gamma-Ray Bursts: The Richness Available for EFOSC2; Monaco, L.; Saviane, I.; Podsiadlowski, P.; Wolf, C.; Gal-Yam, A.; Pasquini, and Beauty of the Global View; Chincarini, G.; 132, 20 L.; Crawford, I.; Mazzali, P.; Pauldrach, A.; Margutti, R.; Covino, S.; D’Avanzo, P.; Fugazza, The ALMA Antenna Transporter; Kraus, M.; Nomoto, K.; Benetti, S.; Cappellaro, E.; Elias- D.; Guidorzi, C.; Mao, J.; Moretti, A.; Capalbi, M.; Stanghellini, S.; Martinez, P.; Koch, F.; Dimmler, Rosa, N.; Hillebrandt, W.; Leonard, D.; Pastorello, Cusumano, G.; D’Elia, V.; Della Valle, M.; Fiore, F.; M.; Moresmau, J. M.; Rykaczewski, H.; 132, 23 A.; Renzini, A.; Sabbadin, F.; Simon, J.; Turatto, Mangano, V.; Molinari, E.; Perri, M.; Romano, P.; Recent Progress at the ALMA Test Facility; Laing, M.; 131, 30 Salvaterra, R.; Zerbi, F.; Campana, S.; Giommi, P.; R.; 132, 28 The Antares Emission Nebula and Mass Loss of Guarneri, A.; Stella, L.; Tagliaferri, G.; Pian, E.; Palazzi, E.; Piranomonte, S.; Antonelli, A.; Salotti, Cute-SCIDAR at Paranal for E-ELT Site alpha Sco A; Reimers, D.; Hagen, H.; Baade, R.; L.; Soto, A. F.; 134, 30 Characterisation; Ramió, H. V.; Reyes, M.; Braun, K.; 132, 33 Delgado, J. M.; Hernández, E.; Cagigal, M. N.; SINFONI Observations of Comet-shaped Knots in The zCOSMOS Data Release 2: the “zCOSMOS- Fuensalida, J. J.; Lombardi, G.; Derie, F.; the Planetary Nebula NGC 7293 (the Helix Neb- bright 10k-sample” and structure in the Universe Navarrete, J.; Sarazin, M.; 132, 29 ula); Matsuura, M.; Speck, A.; Smith, M.; Zijlstra, out to redshifts of order unity; Lilly, S.; The zCOSMOS team; 134, 35 Progress on the European Extremely Large A.; Lowe, K.; Viti, S.; Redman, M.; Wareing, C.; Telescope; Comerón, F.; D’Odorico, S.; Kissler- Lagadec, E.; 132, 37 Patig, M.; Gilmozzi, R.; Spyromilio, J.; 133, 2 Hunting for the Building Blocks of Galaxies like our own Milky Way with FORS; Haehnelt, M. G.; Rauch, M.; Bunker, A.; Becker, G.; Marleau, F.; Graham, J.; Cristiani, S.; Jarvis, M. J.; Lacey, C.; Morris, S.; Peroux, C.; Röttgering, H.; Theuns, T.; 132, 41

64 The Messenger 135 – March 2009 Astronomical News Announcement of the ESO Workshop on Large Award of the Ioannes Marcus Marci Medal to Tom Programmes, 13–15 October 2008, Garching, Wilson, Associate Director for ALMA; Wilson, T.; The 2007 Users Feedback Campaign; Primas, F.; Germany; 132, 50 134, 52 Marteau, S.; Hainaut, O.; Mathys, G.; Romaniello, Announcement of the Topical Symposium “Science Report on the Conference Optical Turbulence M.; Sterzik, M.; 131, 36 with the E-ELT” at JENAM 2008, 8–12 September — Astronomy meets Meteorology; Masciadri, E.; ESO Reflex: A Graphical Workflow Engine for Astro 2008, Vienna, Austria; 132, 51 134, 53 nomical Data Reduction; Hook, R.; Romaniello, Fellows at ESO; 132, 51 Report on the Conference Future Ground-based M.; Ullgrén, M.; Maisala, S.; Solin, O.; Oittinen, T.; ESO Fellowship Programme 2008/2009; 132, 52 Solar System Research: Synergies with Space Savolainen, V.; Järveläinen, P.; Tyynelä, J.; Péron, Probes and Space Telescopes; Käufl, H. U.; Tozzi, New Staff at ESO; 132, 53 M.; Izzo, C.; Ballester, P.; Gabasch, A.; 131, 42 G. P.; 134, 56 Exploring the Cold Universe – A Planetarium Show News from the ESO Science Archive Facility; Report on the ALMA Workshop Simulations for for the IYA 2009; Boffin, H.; Acker, A.; 132, 54 Delmotte, N.; 131, 45 ALMA; Nikolic, B.; Richer, J.; Gueth, F.; Laing, R.; ALMA Science: the ESO–Garching Astronomer’s Personnel Movements; 132, 55 134, 57 View; Testi, L.; 131, 46 Scientific Approach for Optimising Performance, Report on the Conference 400 Years of Astronomi- News from the ALMA Test Facility; Hunter, T.; Laing, Health and Safety in High-Altitude Observatories; cal Telescopes; Brandl, B.; Stuik, R.; 134, 59 Vogt, J.; Nolle-Gösser, T.; Böcker, M.; 133, 49 R.; 131, 47 Announcement of the ESO–Porto Conference Report on the 2007 ESO Fellowship Symposium Report on the ESO and Radionet Workshop on Gas Towards Other Earths: Perspectives and Limita- held at ESO, Vitacura, Chile, 12–14 November and Stars in Galaxies – A Multi-Wavelength 3D tions in the ELT Era; 134, 61 Perspective; De Breuck, C.; Kuntschner, H.; 2007; West, M.; Leibundgut, B.; 131, 48 Announcement of the EIROforum School of Instru- Zwaan, M.; Lehnert, M.; 133, 52 Report on the ESO Chile Science Days held at ESO, mentation (ESI); 134, 61 Vitacura, Chile, 20 November and 5 December ESO at SPIE – Astronomical Telescopes and Instru- ESO ALMA Fellowship Programme; 134, 62 2007; West, M.; 131, 48 ments in Marseille; Moorwood, A.; 133, 57 Daniel Enard 1939–2008; Cullum, M.; 134, 63 Astronomical Observatories and the Republic of In Memoriam Bengt Westerlund; Breysacher, J.; Chile Pave the Way for Future Projects; Argan- Danziger, J.; 133, 58 New Staff at ESO; 134, 64 doña, G.; Mirabel, F.; 131, 49 Do you know your Solar System? Children in Fellows at ESO; 134, 65 Report on the ELSA School on the Science of Gaia Garching do!; Kerber, F.; Kuntschner, H.; ESO at the European City of Science; Boffin, H.; held at the Lorentz Center, Leiden, the Nether- Hanuschik, R.; 133, 58 Heyer, H.; Janssen, E.; 134, 66 lands, 19–28 November 2007; Brown, A.; Lights, Camera, Astronomers! Media training at ESO ESO and the International Year of Astronomy 2009; Lindegren, L.; Kontizas, M.; Turon, C.; Muinonen, Chile; Argandoña, G.; West, M.; 133, 60 Pierce-Price, D.; Russo and Lars Lindberg K.; 131, 50 Social Engagement at ESO; Böcker, M.; 133, 61 Christensen, P.; Lindberg Christensen, L.; 134, 66 Fellows at ESO; 131, 51 New Staff at ESO; 133, 61 Personnel Movements; 134, 67 New Staff at ESO; 131, 52 Fellows at ESO; 133, 63 Announcement of ESO Large Programmes on the Announcement of the Joint ESO, CTIO, ALMA/NRAO Conference Supplement Gran Telescopio Canarias; 131, 53 and Universidad Valparaíso Workshop The Inter- Announcement of the ASTRONET Infrastructure ferometric View on Hot Stars; 133, 64 Special Report on the MPA/ESO/MPE/USM 2008 Roadmap Symposium: An Opportunity to Con- Announcement of the ESO Workshop on ALMA and Joint Astronomy Conference Chemical Evolution tribute to the European Astrophysical Strategy for ELTs: A Deeper, Finer View of the Universe; of Dwarf Galaxies and Stellar Clusters; Primas, F.; the Next 20 Years, 16–19 June 2008, Liverpool, 133, 65 Weiss, A.; 134supp, 2 United Kingdom; 131, 53 Announcement of the Report by the ESA–ESO The Complex Evolution of Simple Systems; Mateo, Announcement of the MPA/ESO/MPE/USM 2008 Working Group on Galactic Populations, M.; 134supp, 3 Joint Astronomy Conference on Chemical Chemistry and Dynamics; 133, 66 Evolution of Dwarf Galaxies and Stellar Clusters; Abundances in Globular Cluster Stars: What is the Personnel Movements; 133, 66 131, 54 Relation with Dwarf Galaxies?; Gratton, R.; Preparing for the ESO Public Surveys with VISTA and 134supp, 9 Announcement of the Joint ESO/INAF-Arcetri VST: New Tools for Phase 2 and a Workshop with Workshop on Future Ground-based Solar System Evidence for Sub-Populations in Globular Clusters: the Survey PIs; Arnaboldi, M.; Dietrich, J.; Research: Synergies with Space Probes and Their Properties and Relationship with Cluster Hatziminaoglou, E.; Hummel, W.; Hussain, G.; Space Telescopes; 131, 55 Properties; Renzini, A.; Milone, A.; Gratton, R.; Neeser, M.; Rejkuba, M.; Bierwirth, T.; Comeron, Cassisi, S.; Bragaglia, A.; Piotto, G.; 134supp, 13 Personnel Movements; 131, 55 F.; Dorigo, D.; Emerson, J.; Nunes, P.; Primas, F.; Linking Chemical Signatures of Globular Clusters to News from the ESO Science Archive Facility; 134, 42 Chemical Evolution; D’Antona, F.; Ventura, P.; Delmotte, N.; 132, 47 Announcement of the Workshop The E-ELT Design 134supp, 18 FP7 E-ELT Preparation – Grant Agreement Funded Reference Mission and Science Plan; 134, 45 Chemical Signatures in Dwarf Galaxies; Hill, V. M.; by the European Commission is Underway; The ESA–ESO Working Group on Galactic Popula- Venn, K. A.; 134supp, 23 Monnet, G.; Gilmozzi, R.; Robinson, M.; 132, 48 tions, Chemistry and Dynamics; Turon, C.; Chemical Evolution of Dwarf Galaxies and Stellar Report on the International Workshop on Star Primas, F.; Binney, J.; Chiappini, C.; Drew, J.; Clusters: Conference Summary; Freeman, K. C.; Formation Across the Milky Way Galaxy; Melo, C.; Helmi, A.; Robin, A.; G. Ryan, S.; 134, 46 134supp, 28 Melnick, J.; Sterzik, M.; 132, 49 Report on the Workshop Interstellar Medium and Star Formation with ALMA: Looking to the Future. A Workshop to Honour Tom Wilson; Martin- Pintado, J.; 134, 50

The Messenger 135 – March 2009 65 Author Index C F

Cassisi, S.; Bragaglia, A.; Gratton, R.; Milone, A.; Freeman, K. C.; Chemical Evolution of Dwarf A Piotto, G.; Renzini, A.; Evidence for Sub-Popula- Galaxies and Stellar Clusters: Conference Sum- tions in Globular Clusters: Their Properties and mary; 134supp, 28 Aldenius, M.; Kerber, F.; Bristow, P.; Nave, G.; Relationship with Cluster Properties; 134supp, 13 Ralchenko, Y.; J. Sansonetti, C.; The Quest for Chincarini, G.; Margutti, R.; Covino, S.; D’Avanzo, P.; Near-infrared Calibration Sources for E-ELT Fugazza, D.; Guidorzi, C.; Mao, J.; Moretti, A.; G Instruments; 133, 14 Capalbi, M.; Cusumano, G.; D’Elia, V.; Della Valle, Argandoña, G.; Mirabel, F.; Astronomical Observato- M.; Fiore, F.; Mangano, V.; Molinari, E.; Perri, M.; Gilmozzi, R.; Monnet, G.; Robinson, M.; FP7 E-ELT ries and the Republic of Chile Pave the Way for Romano, P.; Salvaterra, R.; Zerbi, F.; Campana, Preparation – Grant Agreement Funded by the Future Projects; 131, 49 S.; Giommi, P.; Guarneri, A.; Stella, L.; Tagliaferri, European Commission is Underway; 132, 48 Argandoña, G.; West, M.; Lights, Camera, Astrono- G.; Pian, E.; Palazzi, E.; Piranomonte, S.; Antonelli, Gratton, R.; Abundances in Globular Cluster Stars: mers! Media training at ESO Chile; 133, 60 A.; Salotti, L.; Soto, A. F.; Swift, VLT and Gamma– What is the Relation with Dwarf Galaxies?; Ray Bursts: The Richness and Beauty of the Glo- 134supp, 9 Arnaboldi, M.; Dietrich, J.; Hatziminaoglou, E.; bal View; 134, 30 Hummel, W.; Hussain, G.; Neeser, M.; Rejkuba, M.; Bierwirth, T.; Comeron, F.; Dorigo, D.; Cullum, M.; Daniel Enard 1939–2008; 134, 63 H Emerson, J.; Nunes, P.; Primas, F.; Preparing for the ESO Public Surveys with VISTA and VST: Haehnelt, M. G.; Rauch, M.; Bunker, A.; Becker, G.; New Tools for Phase 2 and a Workshop with the D Marleau, F.; Graham, J.; Cristiani, S.; Jarvis, M. Survey PIs; 134, 42 D’Antona, F.; Ventura, P.; Linking Chemical Signa- J.; Lacey, C.; Morris, S.; Peroux, C.; Röttgering, tures of Globular Clusters to Chemical Evolution; H.; Theuns, T.; Hunting for the Building Blocks of Galaxies like our own Milky Way with FORS; B 134supp, 18 132, 41 D’Odorico, S.; News on the Commissioning of Böcker, M.; Vogt, J.; Nolle-Gösser, T.; Scientific X-shooter; 134, 12 Hook, R.; Romaniello, M.; Ullgrén, M.; Maisala, S.; Solin, O.; Oittinen, T.; Savolainen, V.; Järveläinen, Approach for Optimising Performance, Health and Danziger, J.; Breysacher, J.; In Memoriam Bengt P.; Tyynelä, J.; Péron, M.; Izzo, C.; Ballester, P.; Safety in High-Altitude Observatories; 133, 49 Westerlund; 133, 58 Gabasch, A.; ESO Reflex: A Graphical Workflow Böcker, M.; Social Engagement at ESO; 133, 61 Davies, R.; Rabien, S.; Lidman, C.; Le Louarn, M.; Engine for Astronomical Data Reduction; 131, 42 Bode, M.; Monnet, G.; ASTRONET Roadmap Work- Kasper, M.; Schreiber, N. M. F.; Roccatagliata, V.; Hunter, T.; Laing, R.; News from the ALMA Test ing Group; The ASTRONET Infrastructure Ageorges, N.; Amico, P.; Dumas, C.; Mannucci, F.; Facility; 131, 47 Roadmap: A Twenty Year Strategy for European Laser Guide Star Adaptive Optics without Tip-tilt; Astronomy; 134, 2 131, 7 Boffin, H.; Acker, A.; Exploring the Cold Universe de Zeeuw, T.; The Perfect Machine; 132, 2 K – A Planetarium Show for the IYA 2009; 132, 54 Delmotte, N.; News from the ESO Science Archive Boffin, H.; Janssen, E.; Heyer, H.; ESO at the Facility; 131, 45 Kerber, F.; Hanuschik, R.; Kuntschner, H.; Do you European City of Science; 134, 66 Delmotte, N.; News from the ESO Science Archive know your Solar System? Children in Garching Botticella, M. T.; Cappellaro, E.; Riello, M.; Greggio, Facility; 132, 47 do!; 133, 58 L.; Benetti, S.; Patat, F.; Turatto, M.; Altavilla, G.; Kissler-Patig, M.; Fontana, A.; Venemans, B.; Kneib, Pastorello, A.; Valenti, S.; Zampieri, L.; J.; Doherty, M.; Lidman, C.; Kuntschner, H.; Harutyunyan, A.; Pignata, G.; Taubenberger, S.; E Norris, M.; Larsen, S.; Gieles, M.; Mora Fern- STRESS Counting Supernovae; 134, 25 andes, A.; McCaughrean, M.; Preibisch, T.; Brandl, B.; Stuik, R.; Report on the Conference Eckart, A.; Schödel, R.; García-Marín, M.; Witzel, G.; Seifahrt, A.; Willis, J.; Wehner, E.; Hawk-I – First 400 Years of Astronomical Telescopes; 134, 59 Weiss, A.; Baganoff, F.; Morris, M. R.; Bertram, Results from Science Verification; 132, 7 T.; Dovčiak, M.; Downes, D.; Duschl, W.; Karas, V.; Bristow, P.; Kerber, F.; Rosa, M. R.; Advanced Kraus, M.; Stanghellini, S.; Martinez, P.; Koch, F.; König, S.; Krichbaum, T.; Krips, M.; Kunneriath, Calibration Techniques for Astronomical Spec- Dimmler, M.; Moresmau, J. M.; Rykaczewski, H.; D.; Lu, R.; Markoff, S.; Mauerhan, J.; Meyer, L.; trographs; 131, 2 The ALMA Antenna Transporter; 132, 23 Moultaka, J.; Mužić, K.; Najarro, F.; Pott, J.; Brown, A.; Lindegren, L.; Kontizas, M.; Turon, C.; Schuster, K.; Sjouwerman, L.; Straubmeier, C.; L Muinonen, K.; Report on the ELSA School on the Thum, C.; Vogel, S.; Wiesemeyer, H.; Science of Gaia held at the Lorentz Center, Zamaninasab, M.; Zensus, A.; Probing Sagittarius Laing, R.; Recent Progress at the ALMA Test Facility; Leiden, the Netherlands, 19-28 November 2007; A* and its Environment at the Galactic Centre: 132, 28 131, 50 VLT and APEX Working in Synergy; 133, 26 Lehnert, M.; De Breuck, C.; Kuntschner, H.; Zwaan, Evans, C.; Hunter, I.; Smartt, S.; Lennon, D.; de M.; Report on the ESO and Radionet Workshop Koter, A.; Mokiem, R.; Trundle, C.; Dufton, P.; on Gas and Stars in Galaxies – A Multi- Ryans, R.; Puls, J.; Vink, J.; Herrero, A.; Wavelength 3D Perspective; 133, 52 Simón-Díaz, S.; Langer, N.; Brott, I.; The VLT- Lilly, S.; The zCOSMOS team; The zCOSMOS Data FLAMES Survey of Massive Stars; 131, 25 Release 2: the “zCOSMOS-bright 10k-sample” and structure in the Universe out to redshifts of order unity; 134, 35 Liske, J.; Grazian, A.; Vanzella, E.; Dessauges, M.; Viel, M.; Pasquini, L.; Haehnelt, M.; Cristiani, S.; Pepe, F.; Bonifacio, P.; Bouchy, F.; D’Odorico, S.; D’Odorico, V.; Levshakov, S.; Lovis, C.; Mayor, M.; Molaro, P.; Moscardini, L.; Murphy, M.; Queloz, D.; Udry, S.; Wiklind, T.; Zucker, S.; E-ELT and the Cosmic Expansion History – A Far Stretch?; 133, 10

66 The Messenger 135 – March 2009 M P T

Martin-Pintado, J.; Report on the Workshop Patat, F.; Chandra, P.; Chevalier, R.; Justham, S.; Testi, L.; ALMA Science: the ESO–Garching Interstellar Medium and Star Formation with Podsiadlowski, P.; Wolf, C.; Gal-Yam, A.; Pasquini, Astronomers View; 131, 46 ALMA: Looking to the Future A Workshop to Hon- L.; Crawford, I.; Mazzali, P.; Pauldrach, A.; Turon, C.; Primas, F.; Binney, J.; Chiappini, C.; Drew, our Tom Wilson; 134, 50 Nomoto, K.; Benetti, S.; Cappellaro, E.; Elias- J.; Helmi, A.; Robin, A.; G. Ryan, S.; The ESA– Masciadri, E.; Report on the Conference Optical Rosa, N.; Hillebrandt, W.; Leonard, D.; Pastorello, ESO Working Group on Galactic Populations, Turbulence — Astronomy meets Meteorology; A.; Renzini, A.; Sabbadin, F.; Simon, J.; Turatto, Chemistry and Dynamics; 134, 46 134, 53 M.; Seeking for the Progenitors of Type Ia Super- novae; 131, 30 Mateo, M.; The Complex Evolution of Simple Systems; 134supp, 3 Pierce-Price, D.; Russo and Lars Lindberg U Christensen, P.; Lindberg Christensen, L.; ESO Matsuura, M.; Speck, A.; Smith, M.; Zijlstra, A.; and the International Year of Astronomy 2009; Käufl, H. U.; Tozzi, G. P.; Report on the Conference Lowe, K.; Viti, S.; Redman, M.; Wareing, C.; 134, 66 Future Ground-based Solar System Research: Lagadec, E.; SINFONI Observations of Comet- Synergies with Space Probes and Space Tele- shaped Knots in the Planetary Nebula NGC 7293 Primas, F.; Marteau, S.; Hainaut, O.; Mathys, G.; scopes; 134, 56 (the Helix Nebula); 132, 37 Romaniello, M.; Sterzik, M.; The 2007 Users Feedback Campaign; 131, 36 McMahon, R.; Parry, I.; Venemans, B.; King, D.; Primas, F.; Weiss, A.; Special Report on the MPA/ Ryan-Weber, E.; Bland-Hawthorn, J.; Horton, A.; V DAZLE on the VLT; 131, 11 ESO/MPE/USM 2008 Joint Astronomy Confer- ence Chemical Evolution of Dwarf Galaxies and Meisenheimer, K.; Raban, D.; Tristram, K.; van Belle, G. T.; Sahlmann, J.; Abuter, R.; Accardo, Stellar Clusters; 134supp, 2 Schartmann, M.; Jaffe, W.; Röttgering, H.; M.; Andolfato, L.; Brillant, S.; de Jong, J.; Derie, F.; Burtscher, L.; Mid-infrared Interferometry of Delplancke, F.; Duc, T. P.; Dupuy, C.; Gilli, B.; Active Galactic Nuclei: an Outstanding Scientific Gitton, P.; Haguenauer, P.; Jocou, L.; Jost, A.; Di R Success of the VLTI; 133, 36 Lieto, N.; Frahm, R.; Ménardi, S.; Morel, S.; Moresmau, J.; Palsa, R.; Popovic, D.; Pozna, E.; Melo, C.; Pasquini, L.; Downing, M.; Deiries, S.; Ramió, H. V.; Reyes, M.; Delgado, J. M.; Hernández, Puech, F.; Lévêque, S.; Ramirez, A.; Schuhler, N.; Naef, D.; Hanuschik, R.; Palsa, R.; Castillo, R.; E.; Cagigal, M. N.; Fuensalida, J. J.; Lombardi, G.; Somboli, F.; Wehner, S.; ESPRI Consortium, T.; Peña, E.; Bendek, E.; Gieles, M.; Detector Derie, F.; Navarrete, J.; Sarazin, M.; Cute-SCIDAR The VLTI PRIMA Facility; 134, 6 Upgrade for FLAMES: GIRAFFE Gets Red Eyes; at Paranal for E-ELT Site Characterisation; 132, 29 133, 17 van den Ancker, M.; Fedele, D.; Petr-Gotzens, M.; Reimers, D.; Hagen, H.; Baade, R.; Braun, K.; The Rafanelli, P.; A Multi-Wavelength Study of the Minniti, D.; Melo, C.; Naef, D.; Udalski, A.; Pont, F.; Antares Emission Nebula and Mass Loss of alpha 2003–2006 Outburst of V1647 Orionis; 131, 20 Moutou, C.; Santos, N.; Queloz, D.; Mazeh, T.; Sco A; 132, 33 Gillon, M.; Mayor, M.; Udry, S.; Diaz, R.; Hoyer, S.; Venn, K. A.; Hill, V. M.; Chemical Signatures in Ramirez, S.; Pietrzynski, G.; Gieren, W.; Ruiz, M. Dwarf Galaxies; 134supp, 23 T.; Zoccali, M.; Tamuz, O.; Shporer, A.; Kubiak, S M.; Soszynski, I.; Szewczyk, O.; Szymanski, M.; Ulaczyk, K.; Wyrzykowski, L.; Behind the Scenes Sarazin, M.; Melnick, J.; Navarrete, J.; Lombardi, G.; W of the Discovery of Two Extrasolar Planets: ESO Seeing is Believing: New Facts about the Evolu- Large Programme 666; 133, 21 tion of Seeing on Paranal; 132, 11 West, M.; Leibundgut, B.; Report on the 2007 ESO Fellowship Symposium held at ESO, Vitacura, Monnet, G.; Report on the JENAM 2008 Meeting Saviane, I.; Monaco, L.; Two Volume-phased Chile, 12–14 November 2007; 131, 48 Symposium Science with the E-ELT; 134, 14 Holographic Grisms Now Available for EFOSC2; Moorwood, A.; ESO at SPIE – Astronomical 132, 20 West, M.; Report on the ESO Chile Science Days held at ESO, Vitacura, Chile, 20 November and 5 Telescopes and Instruments in Marseille; 133, 57 Snodgrass, C.; Saviane, I.; Monaco, L.; Sinclaire, P.; December 2007; 131, 48 Morelli, L.; Pompei, E.; Pizzella, A.; Méndez-Abreu, EFOSC2 Episode IV: A New Hope; 132, 18 J.; Maria Corsini, E.; Coccato, L.; Saglia, R.; Sarzi, Spyromilio, J.; Comerón, F.; D’Odorico, S.; Kissler- M.; Bertola, F.; Stellar Populations of Bulges of Patig, M.; Gilmozzi, R.; Progress on the European Disc Galaxies in Clusters; 133, 31 Extremely Large Telescope; 133, 2 Sterzik, M.; Melnick, J.; Melo, C.; Report on the International Workshop on Star Formation Across N the Milky Way Galaxy; 132, 49 Nardetto, N.; Kervella, P.; Barnes, T.; Bersier, D.; Sullivan, M.; Balland, C.; The Supernova Legacy Fokin, A.; Fouqué, P.; Gillet, D.; Groh, J.; Kraus, Survey; 133, 42 S.; Mathias, P.; Mérand, A.; Millour, F.; Mourard, D.; Stoekl, A.; From the Dynamics of Cepheids to the Milky Way Rotation and the Calibration of the Distance Scale; 134, 20 Nikolic, B.; Richer, J.; Hills, R.; Stirling, A.; Phase Correction for ALMA: Adaptive Optics in the Submillimetre; 131, 14 Nikolic, B.; Richer, J.; Gueth, F.; Laing, R.; Report on the ALMA Workshop Simulations for ALMA; 134, 57

The Messenger 135 – March 2009 67 ESO is the European Organisation for Contents Astronomical Research in the Southern Hemisphere. Whilst the Headquarters The Organisation (comprising the scientific, technical and R. Fischer, J. Walsh — An Extension for ESO Headquarters 2 administrative centre of the organisa- tion) are located in Garching near Telescopes and Instrumentation ­Munich, Germany, ESO operates three A. Baudry — The ALMA Correlator: Performance and observational sites in the Chilean Ata­- Science Impact in the Millimetre/Submillimetre 6 cama desert. The Very Large Telescope M. Scodeggio et al. — Improving the Multiplexing of VIMOS MOS (VLT), is located on Paranal, a 2 600 m Observations for Future Spectroscopic Surveys 13 high mountain south of Antofagasta. At C. Melo et al. — Six Years of FLAMES Operations 17 La Silla, 600 km north of Santiago de Chile at 2 400 m altitude, ESO operates Astronomical Science several medium-sized optical tele­ S. Hubrig et al. — Studying the Magnetic Properties of scopes. The third site is the 5 000 m Upper Main-sequence Stars with FORS1 21 high Llano de Chajnantor, near San F. Millour et al. — Wolf-Rayet Stars at the Highest Angular Resolution 26 Pedro de Atacama. Here a new submil- A. Richichi et al. — The Beauty of Speed 32 limetre telescope (APEX) is in opera- H. Horst et al. — VISIR Observations of Local Seyfert Nuclei and tion, and a giant array of submillimetre the Mid-infrared — Hard X-ray Correlation 37 antennas (ALMA) is under development. J. Kurk et al. — A VLT Large Programme to Study Galaxies at z ~ 2: Over 2 000 proposals are made each GMASS — the Galaxy Mass Assembly Ultra-deep Spectroscopic Survey 40 year for the use of the ESO telescopes. Astronomical News The ESO Messenger is published four E. Hatziminaoglou, F. Chéreau — VirGO: A Visual Browser for the ESO times a year: normally in March, June, Science Archive Facility 46 September and December. ESO also N. Delmotte — News from the ESO Science Archive Facility 49 publishes Conference Proceedings and P. Garcia — New Infrastructures Require New Training: other material connected to its activi- The Example of the Very Large Telescope Interferometer Schools 50 ties. Press Releases inform the media G. Mathys, B. Leibundgut — Large Programmes 53 about particular events. For further P. Russo et al. — ESO and the International Year of Astronomy 2009 in­formation, contact the ESO Education Opening Ceremony 54 and Public Outreach Department at the MAD and Beyond: Science with Multi-Conjugate Adaptive following address: Optics Instruments 56 IR and Sub-mm Spectroscopy: A New Tool for Studying Stellar Evolution 56 ESO Headquarters Detectors for Astronomy 2009 57 Karl-Schwarzschild-Straße 2 M. West et al. — ESO’s Studentship Programmes: 85748 Garching bei München Training Tomorrow’s Astronomers Today 57 Germany ESO Studentship Programme 59 Phone +498932006-0 New Staff at ESO 60 Fax +49893202362 Fellows at ESO 62 [email protected] Personnel Movements 63 www.eso.org Annual Index 2008 (Nos. 131–134) 64

The ESO Messenger: Editor: Jeremy R. Walsh Layout/typesetting: Mafalda Martins Design: Jutta Boxheimer www.eso.org/messenger/

Printed by Peschke Druck Front Cover: Colour composite of the centre of Schatzbogen 35 NGC 5128 containing the radio source Centaurus A. 81805 München Three separate image sets are combined: APEX Germany LABOCA 870 μm image in orange; Chandra X-ray Observatory combined 0.5–2 keV image in blue; © ESO 2009 visible light (B, V and I band) WFI images from the ISSN 0722-6691 MPG/ESO 2.2 m telescope. See ESO 03/09 for details.