1. Brasil to join ESOThe 2nd generation VLTI instrument GRAVITYSpectroscopy of planet-forming discsLarge Lyman-break galaxy surveyThe Messenger No. 143 March 2011

2. The OrganisationAdriaan Blaauw, 19142010In the last issue of The MessengerThere follow three tributes to Adriaan Pottasch; and by Raymond Wilson, who(142, p. 51) only a brief obituary of Adriaan Blaauw: by Tim de Zeeuw, current led the Optics Group during his tenure asBlaauw, the second Director General ESO Director General; by his long-term Director General.of ESO, could be included at the time ofcolleague at the Kapteyn Institute, Stuartgoing to press.Tim de Zeeuw1 but moved to Yerkes Observatory inESO [M]1953, becoming its associate director in1956, and moved back to Groningen1ESO in 1957, where he was in a key position tocontribute to transforming the idea ofBaade and Oort into reality. He was Sec-Professor Adriaan Blaauw, ESOs sec-retary of the ESO Committee (the proto-ond Director General and one of the Council) from 1959 through 1963, amost influential astronomers of the twen- period which included the signing of thetieth century, passed away on 1 Decem-ESO Convention on 5 October 1962.ber 2010. Blaauw became ESOs Scientific Directorin 1968. In this position he also pro-Adriaan Blaauw was born in Amsterdam, vided the decisive push which led to thethe Netherlands, on 12 April 1914. He creation of Astronomy and Astrophysics,studied astronomy at Leiden University, which successfully combined andunder de Sitter, Hertzsprung and Oort,replaced the various individual nationaland obtained his doctorate (cum laude)journals for astronomy, and today iswith van Rhijn at the Kapteyn Laboratoryone of the leading astronomy researchin Groningen in 1946. His PhD thesispublications in the world. The articlewas entitled A study of the Scorpio by Pottasch (1994) and the following trib-Centaurus Cluster. During his career,ute provide further details of BlaauwsBlaauw became renowned for his ground-creative leadership in the founding of thebreaking studies of the properties of European astronomical journal.OB associations (groups of young, hotstars) which contain the fossil imprint Blaauw was Director General from 1970Figure 1. Adriaan Blaauw in 1973 while Director Gen-through 1974. During this period several eral of ESO. From a photograph taken during a con-of their star formation history. Perhaps his tract-signing ceremony for building works at La Silla.most famous work explained why some telescopes, including the ESO 0.5-metreOB stars are found in isolation travellingand 1-metre Schmidt telescopes, beganat unusually high velocity: the so-called operating at ESOs first observatory The Messenger may serve to give therun-away stars. Blaauw proposed insite, La Silla, in Chile, and much workworld outside some impression of what1961 that these stars had originally been was done on the design and constructionhappens inside ESO. The continuingmembers of binary systems, and when of the ESO 3.6-metre telescope, whichpopularity of The Messenger is a testi-one star in the binary experiences ahad its first light in 1976. Blaauw decidedmony to Blaauws foresight.supernova explosion, its companion sud- that it was crucial for this project to movedenly ceases to feel the gravitationalESOs Headquarters and the Technical After stepping down as Director Generalpull that keeps it in its orbit and hence itDepartment from Hamburg to Geneva, toof ESO, Blaauw returned to Leiden,runs away at its orbital velocity.benefit from the presence of the experi- where I had the privilege to be amongstenced CERN engineering group. He alsohis students. He continued to play aIn addition to his distinguished research oversaw the development of the Proto-very important role in international astron-career, Blaauw played a central role in col for Privileges and Immunities that isomy. He was President of the Interna-the creation of ESO. In 1953, Baade and critical for ESOs functioning. In May tional Astronomical Union from 1976 toOort proposed the idea of combining 1974 he launched The Messenger with1979, during which period he used hisEuropean resources to create an astro-the stated goal: to promote the partici-considerable diplomatic skills to convincenomical research organisation thatpation of ESO staff in what goes on in China to rejoin the IAU. From 1979 tocould compete in the international arena. the organisation, especially at places of1982 he served on the ESO Council onBlaauw had returned to Leiden in 1948,duty other than our own. Moreover, behalf of the Netherlands. He retired from2 The Messenger 143 March 2011 3. his Leiden professorship in 1981 and ESOs early history with some of us of science, honorary doctorates frommoved back to Groningen, but stayed(see the photograph in The Messenger, the University of Besancon and fromactive in various areas. This included 137, p.6). During this visit he revealed his lObservatoire de Paris and, like his pre-organising the historical archives of ESOwish to visit Chile one more time if hisdecessor as ESO Director General,and of the IAU work which resulted inhealth would allow this. It was a pleasureOtto Heckman, the Bruce Medal of thetwo books, ESOs Early History (Blaauw,to organise this trip in February 2010. Astronomical Society of the Pacific. He1991) and History of the IAU (Blaauw,He met ESO legends Albert Bosker, Jan was well known for his warm personality,1994). He also served as Chairman of Doornenbal, Erich Schumann and Daniel wisdom, humour, legendary patience,the Scientific Evaluation Committee forHofstadt and was driven to La Silla and and the rare gift of being able to slowthe European Space Agency satelliteParanal by car to enjoy Chiles beautiful down when the pressure mounted. TheHIPPARCOS, advising on many aspectslandscapes. He characteristically engaged personal account of his life, entitledof its scientific programme. When theyoung people at the telescopes andMy Cruise Through the World of Astron-data became available in 1996, he wasin Vitacura in interesting discussions andomy, published in the 2004 Annualactively involved in the re-analysis of thethroughout the visit displayed a crystal- Reviews of Astronomy and Astrophysicsyoung stellar groups that he had studied clear perspective on the development of (Blaauw, 2004), provides an accuratefirst during his PhD research. ESO and on the exciting opportunities and inspiring picture of a truly remarkable for the future programme (a photographperson, who positively influenced theBlaauw remained keenly interested in of this visit is shown in The Messenger,lives of many.developments at ESO. After a discussion139, p. 61). The characteristic twinkle inwith him in late 2008, he drove himselfhis eye was as bright as ever.to Garching and back in July 2009 in Referencesorder to take another look at the historical Blaauw won many academic distinctions,Blaauw, A. 2004, ARAA, 42, 1documents in the library and to discussincluding membership of many academiesPottasch, S. R. 1994, The Messenger, 76, 62Stuart Pottasch1 tronomers of PhD level or higher, with theThis is where Adriaan, who was at that result that 75 % of those present agreedtime Scientific Director of ESO, came that a new journal was desirable. Simi- in. He suggested, organised and imple-1Kapteyn Laboratorium, Groningen, lar meetings took place at a somewhat mented a legal status for the new jour-the Netherlandshigher level in other countries. At thisnal. The basic idea was that ESO would point there was much enthusiasm to beginmake use of the fact that it was an offi- a new journal. This led to a meeting of cial European organisation. Its adminis-Adriaan has contributed to many fields ofEuropean astronomers on 8 April 1968. trative and legal services were madeastronomy. In the long years we have available to the journal through a formalknown and worked with each other there In spite of the enthusiasm for the Euro-agreement between ESO and the Boardare two aspects that may be less wellpean astronomical journal, there were of Directors of the journal. This agree-known and that I would like to highlight.rather difficult problems ahead. Thesement was confirmed at the December problems were of a practical nature and 1968 ESO Council meeting, just beforeFirst of all is the deep interest he took in arose because the new journal was tothe first issue of the new journal Asthe formation of the European journalbe a combination of journals published in tronomy and Astrophysics appeared inAstronomy and Astrophysics. Adriaanvarious European countries. The indi- January 1969. Individual countriestook part in the initial discussions, whichvidual journals all had a rather differentcould now contribute financially to thefirst began to take real shape in 1967 status. Some were owned by privatejournal, but ESO itself would carry noand especially in 1968. The discussionspublishers, some by astronomical organi-financial responsibility for the journal. Atin 1967 took place in several European sations. The French journals were owned the same time the Board would be en-countries. At first they were independentby the ministry in France, which couldtirely independent of any influence fromof each other and took place because not contribute financially to a Europeanthe ESO side on its scientific policy.of a general feeling in Europe that existing journal without an official treaty betweenEuropean astronomical journals werevarious countries. The timescale forBut this did not end Adriaans connec-not being read to the same extent as the such a treaty, essentially the creation oftion with the new journal. He acceptedAmerican journals. In December 1967an international organisation, was ex-an invitation to become a member ofa meeting took place in France which pected to be long, and the discussionsthe Board of Directors and was in factwas attended by almost all French as-complicated.elected chairman of that body. The The Messenger 143 March 2011 3 4. The Organisation de Zeeuw T., Pottasch S., Wilson R., Adriaan Blaauw, 19142010importance of this can be seen in the fact where Adriaan was able to reconcile theto combine his scientific curiosity withthat the journal at the time was moredifferences. He was chairman of the A&Avarious administrative responsibilitiesturbulent than it is at present. Not onlyBoard for about ten years. without letting the one cloud out thewere there more disputes between indi-other. I think that he was able to do thisvidual scientists, there were also dis-A second aspect of Adriaans careerbecause he approached science inputes between different countries, espe- that is worth highlighting can be stated an unhurried and patient way. Astron-cially about the refereeing. Some of more simply. He remained an active omy interested him; there was alwaysthese disputes were brought to the Board scientist for his whole life, and was able time for it.Raymond Wilson1building on the CERN campus. The total would have been no active optics at ESO staff in this fledgling technical division and, consequently, no NTT, VLT or E-ELT of ESO cannot have numbered more thanproject. The readers of this tribute will1Rohrbach/Ilm, Germanyten or twelve. understand, I am sure, why I hold AdriaanBlaauw in such high esteem. A major contractual problem now emerged.It is an honour and a pleasure to write aI had clearly understood, from BlaauwsFinally, there was another aspect oftribute to Adriaan Blaauw, whom I con- interview with me, that I would be the his leadership which I greatly admired.sider to be an underrated Director Gen-leader of a newly-founded Optics Group,Once settled in with my new Opticseral of ESO, above all through being indealing with all optical aspects of tele-Group, things were going quite well forthe long shadow thrown by his successorscopes (at that time, mainly the 3.6-metre me and I was elected to be Staff Rep-Lodewijk Woltjer.telescope) and instrumentation. How- resentative. In Blaauws weekly one-day ever, in the technical group, led by Svend visits to Geneva, I was always the firstI am unable to make any comments Laustsen, the responsibility for telescope person he visited. But he was not con-regarding his achievements in the astro- optics was in the hands of a Germancerned about my technical function,nomical field. I am only going to commentastronomer, Alfred Behr, and for instru- which we had organised: he left that toon my personal experience of his workmentation optics in the hands of AndersLaustsen, who had, of course, acceptedas ESO Director General, above all at theReiz, a Danish astronomer. My role inthe new Optics Group, in which Behrstime when I was engaged by him per-this existing structure appeared only to work was now integrated under my lead-sonally to create and head a new Opticsbe that of a senior assistant to them, ership. No, he visited me first as StaffGroup on the technical side of ESOs above all to Alfred Behr. This situation Representative to ask if the staff wereactivities. At this time, his office was still was unacceptable to me and not as I hadcontent or whether there were any prob-in Hamburg, where ESO was founded, understood the scope of the position I lems where he should intervene. Thisabove all, by Professor Otto Heckman,had accepted.proves again his absolutely fair and hu-for the 3.6-metre telescope project. This mane leadership!project was intended to bring ESO up toBlaauw normally only came to Genevathe level of the American telescopes with, for one day a week. However, when IAdriaan Blaauw was not only a great ESOat that time, one of the larger telescopes rang him up and explained the gravity of Director General, he was also an admi-built in the post-Palomar (5-metre) era. the situation and the inevitability of myrable gentleman of impeccable integrity. leaving ESO immediately if he could notI left the firm of Carl Zeiss to go to ESO rectify it, he came at once and we dis-in 1972, when Zeiss, at the time of acussed the matter over another goodserious recession in German industry,lunch. I emphasised my clear position onstarted laying off staff, including those of the matter and that I would try to returnmy own Optical Design department,to Zeiss immediately, in spite of the badwhere I had conceived my idea of activesituation there. Blaauw recognised thatoptics. Professor Blaauw interviewed I was very serious and stated he wouldme over a good lunch in Geneva. He inform Laustsen at once that a new Opticsimmediately offered me a senior position Group would immediately be foundedat ESO in Geneva, where, through his under my leadership. Without this boldinitiative, ESO had a small barrack-type and clear direction by Blaauw there4 The Messenger 143 March 2011 5. The OrganisationBrazil to Join ESOTim de Zeeuw1 now be submitted to the Brazilian Par-nities for Brazilian high-tech industry toliament for ratification. The signing ofcontribute to the ESO programme, in-the agreement followed its unanimouscluding the European Extremely Large1ESO approval by the ESO Council during an Telescope project. It will also bringextraordinary meeting, by teleconference, new resources and skills to the organi-on 21 December 2010.sation at the right time for them to makeOn 29 December 2010, at a ceremonya major contribution to this exciting pro-in Brasilia, the Brazilian Minister of Sci- Joining ESO will give new impetus to the ject, added Tim de Zeeuw.ence and Technology, Sergio Machado development of science, technology andRezende and the ESO Director General, innovation in Brazil as part of the consid- The president of ESOs governing body,Tim de Zeeuw signed the formal acces- erable efforts our government is making the Council, Laurent Vigroux, concluded:sion agreement, paving the way for Brazil to keep the country advancing in theseAstronomers in Brazil will benefit fromto become a Member State of the Euro- strategic areas, said Minister Rezende.collaborating with European colleagues,pean Southern Observatory. Brazil willand naturally from having observing timebecome the fifteenth Member State and The membership of Brazil will give the at ESOs world-class observatories atthe first from outside Europe. Since thevibrant Brazilian astronomical communityLa Silla, Paranal and APEX at Chajnantor,agreement implies accession to an inter-full access to the most productive obser- as well as on ALMA, which ESO is con-national convention, the agreement must vatory in the world and open up opportu-structing with its international partners.Figure 1. ESO Director General, Tim de Zeeuw,(right) in discussion with the Brazilian Ministerof Science and Technology, Sergio MachadoRezende, during the accession ceremony inBrasilia on 29 December 2010. The Messenger 143 March 20115 6. Telescopes and Instrumentation Fisheye image of the interior of the dome for VLT UT4 Yepun. See potw1049 for details. HH 30 2MASSWJ1207334-393254778 mas55 AU at 70pc N200 AUE Detection of intermediate Ten year largemass BH in GCs/Arches programme Orbit of exo- Stellar motions in nuclei Jupiter/UranusDetection of SR/GR effectsof nearby galaxies in cusp star orbits Three year largeAstrometric signalDetection of dark halo around SgrA* programme exo-Jupiter/Uranus3D dynamics of nuclear star clusterEvolution outows in Gas ows in AGNYSOs & micro-QSOs SgrA* are dynamics Single season Proper motions massive star cluster campaignImaging jets/discs in YSOs & CBs Binary dynamicsLensing10 0 10 2 10 410 6Maximum distance from Earth (pc)Key experiments with GRAVITY are illustrated (seearticle by Eisenhauer et al. p. 16). Clockwise fromS27 S31 S19 S12top left are: jet/discs in a nearby star-forming S29region; planet-brown dwarf binary; dust disc with S5S14 S17S4 S2 central gap; Arches star cluster; M31 star discs;S6NGC1068 outflow/narrow line region; modelling ofS39a Galactic Centre flare; radial precession of stellarS21 orbits; S-star orbits; nuclear star cluster and radioS1S13 S18emission in the Galactic Centre. In the central S8 S33inset the horizontal axis denotes the maximum S9S24distance from Earth, the vertical axis the time spanof the measurements.6 The Messenger 143 March 2011 7. Telescopes and InstrumentationHARPSpol The New Polarimetric Mode for HARPSNikolai Piskunov1and linear polarisations across their pro-This sets very stringent limits on the di-Frans Snik 2 files. For non-degenerate objects, themensions of the polarimeter, because itAndrey Dolgopolov 3continuum is mostly unpolarised, whichneeds to fit in between various mecha-Oleg Kochukhov1offers a reliable intrinsic calibration that is nisms (calibration light feeds, calibrationMichiel Rodenhuis 2necessary for measuring very weak fields, mirror and fibre cover) filling the adapter.Jeff Valenti 4 but such measurements require a veryThe polarimeter consists of the enclo-Sandra Jeffers 2 stable spectropolarimetric instrument.sure hosting a precision horizontal slider.Vitaly Makaganiuk1 The slider holds two identical opticalChristopher Johns-Krull 5The HARPS spectrograph at ESOs tables installed perpendicular to the slid-Eric Stempels1 3.6-metre telescope at La Silla is one of ing direction. Each optical table containsChristoph Keller 2 the most successful spectroscopic as- a full set of polarisation optics (Figure 1), tronomical instruments ever built (Mayorseparating the incoming light into two et al., 2003). The exceptional temporal beams. Since the polarising beam-splitter1Department of Physics and Astronomy, and spatial stability of HARPS makesposition is fixed relative to the fibres, theUppsala University, Sweden it an ideal instrument for spectropolarim-polarisation of the incoming light needs2Sterrekundig Instituut Utrecht, Utrechtetry. The new polarimeter takes full ad-to be converted to the frame of the beam-University, the Netherlandsvantage of the two optical fibres to bringsplitter. This is achieved by rotating wave3Crimean Astrophysical Observatory, the collected light, split into two orthogo-plates in front of the beam-splitters: aCrimea, Ukrainenal polarisations, from the Cassegrainhalf-wave plate for the linear polarimeter4STScI, Baltimore, USAfocus of the 3.6-metre telescope to the and a quarter-wave plate for the circular5Rice University, Houston, USAHARPS spectrograph. Analysing polari- one. The relative intensity of the two sations at the Cassegrain focus mini- beams at each wavelength carries the mises the influence of instrumentation on information about the polarisation of theThe HARPS spectrograph can now the measurements. The new module, light.perform a full polarisation analysis ofcalled HARPSpol, allows sensitive andspectra. It has been equipped with accurate measurements of both circularThe polarising beam-splitters consista polarimetric unit, HARPSpol, which and linear polarisations of stellar light of a Foster prism (a modified Glanwas jointly designed and producedas a function of wavelength, at high spec-Thompson polariser). The primary beamby Uppsala, Utrecht and Rice Univer- tral resolution. In this article we give asuffers from crystal astigmatism, whichsities and by the STScI. Here we pre-short presentation of the polarimeter and is corrected by a cylindrical lens. Thesent the new instrument, demonstrate show some results from the first year ofsecondary beam is deviated by 45.its polarisation capabilities and show operation.Beam-channelling prisms align the opti-the first scientific results.cal axis and the focus of the secondary beam with the second HARPS fibre. HARPSpol Whats inside the box? The selected optical scheme solves twoIntroduction HARPSpol is installed inside the Casse- Figure 1. Schematic of the HARPSpol optical design.Spectropolarimetry is one of a very fewgrain adapter, located directly below the Left: the view in the sliding direction. Right: side viewdirect ways of detecting and studyingprimary mirror of the 3.6-metre telescope.of the two polarimeters.magnetic fields. Magnetic fields are pre-sumed to play crucial roles in all kindsof objects and environments in space,stirring turbulence, transporting angularmomentum, converting kinetic energyto radiation, controlling plasma motion,etc. Magnetic fields create polarisation inspectral lines though the Zeeman effect,and thus polarisation measurementsallow us to measure the strength and theorientation of the field vector, providingimportant clues for understanding starformation, the origin of structures in stel-lar atmospheres and stellar activity. Infact, the origin and the evolution of mag-netic fields remains one of the mostimportant topics in modern astrophysics.Spectral lines formed in the presence ofa magnetic field generally exhibit circular The Messenger 143 March 2011 7 8. Telescopes and InstrumentationPiskunov N. et al., HARPSpol The New Polarimetric Mode for HARPSdifficulties: (1) it is highly achromatic,Figure 2. HARPSpolis shown during installa-that is, the image of a star after projectiontion. The HARPSpolthrough HARPSpol is essentially the enclosure is on the right.same in the red and in the blue parts ofThe slider is in the linearthe spectrum; and (2) slight errors inpolarisation position.The half-wave plate forpositioning of the slider do not affect thethe linear polarimeteroptical/polarisation performance. Moreis visible in the middle ofinformation about the optical designthe picture. The roundof HARPSpol can be found in Snik et al. mirror below the linearpolarimeter is one of the(2008, 2010).HARPS fibre heads.The selected wave plates are super-achromatic. They consist of five layers ofbirefringent polymer. This makes thepolarimeters suitable for the entire HARPSwavelength range (380690 nm) withoutintroducing (polarised) fringes. The simul-taneous measurements in two polari-sation directions, together with the polari-sation modulation by the wave plates,renders the polarimetry with HARPSpolto first order insensitive to seeing andfibre/spectrograph throughput (Semel etal., 1993; Bagnulo et al., 2009).IntegrationOnce installed at the Cassegrain adapter,HARPSpol was integrated with the HARPSinstrument control electronics and soft-ware. When inserted into the optical path,HARPSpol shifts the focus of the tele-scope by approximately 2 mm, which iscompensated for by moving the sec-ondary mirror. Figure 2 shows HARPSpolinstalled inside the Cassegrain adapter.Spectropolarimetry is performed byFigure 3. The total throughput from the telescope tothe detector with and without HARPSpol is shown.selecting the corresponding template(s)The sharp drops are not real: they are due to hydro-in the observing software. Calibration andgen lines that are treated differently in spectro-science templates are available for cir-photometry and spectropolarimetry.cular and linear polarimetry. The scienceFigure 4. The combined averageprofile for intensity and polarisation(lower and middle plots) for Cen A.Left panel shows circular polarisationmeasurements (Stokes parameter V).Middle and right panels are for linearpolarisations. The null profile is shownuppermost. 8 The Messenger 143 March 2011 9. Figure 5. Comparison of the Stokesspectra of a standard magnetic star Equ taken at the CFHT with theESPANDONS spectropolarimeter (redline) and with HARPSpol (black line) isshown. The ESPADONS spectra weretaken as part of CFHTs calibrationand engineering plan, and wereretrieved from the Canadian Astron-omy Data Centre. The visible differ-ences are mostly due to the higherresolving power of HARPS.Figure 6. One of the HARPS polarisa-tion spectra of a CP star, HD 24712,is shown. Both circular and linearpolarisations are detected for practi-cally every spectral line.template allows multiple exposures to to the lower throughput of the sky fibre plot in each panel of Figure 4 shows thebe taken in the selected mode (circular (used to carry one of the polarisedso-called null spectrum, obtained byor linear) for a sequence of wave-plate beams), but still sufficient to reach rather modifying the analysis in such a way asangles. The full complement of polarisa-faint targets. to destroy the polarisation signal in thetion characteristics can be registered incoming light (Bagnulo et al., 2009).in six or twelve exposures, with the latter Systematic errors limit both the polari- What remains reflects the spurious polari-offering intrinsic control over spuriousmetric sensitivity and the accuracy. The sation induced inside the instrumentationpolarisation signals. The HARPSpol pipe-sensitivity is the weakest polarisationor by the data reduction.line then processes the data and thedetectable with HARPSpol. After accu-final products include the Stokes param-mulating enough photons we expectWe do not expect any detectable polari-eters as a function of wavelength.to see spurious polarisation present insation signal from Cen A and Figure 4the light coming to the telescope. We test shows that our new instrument does notthis by observing a bright source anddetect or induce any polarisation aboveHARPSpol: Performance collecting many photons in a series of the level of 10 5, which is on a par withmany short exposures. Figure 4 shows the best solar polarimeters like ZIMPOLDuring commissioning we have meas-the results of the test for an inactive solar- (Ramelli et al., 2010). The accuracy (theured several characteristics of HARPSpol. type star, Cen A, where we reach the level at which the HARPSpol measure-The most important ones for the ob- median signal-to-noise ratio of 2 400 perments match the true polarisation signal)server are the total throughput of the sys- CCD column. Besides combining multi- is assessed by observing objects withtem and the polarimetric sensitivity. The ple exposures we also derive the meanknown polarisation spectra. Our observa-throughput (Figure 3) was measuredStokes profiles using the least squarestions of Equ demonstrate the highby observing spectrophotometric stand-deconvolution (LSD) technique (Donati et accuracy of HARPSpol. Equ is a well-ards, reducing the data, rebinning it al., 1997; Kochukhov et al., 2010), whichstudied magnetic star showing linear andto match the resolution of the spectro- takes advantage of the fact that mostcircular polarisations. The lack of notice-photometry and deriving the sensitivity of the spectral lines are affected by mag- able rotation makes Equ an excellentcurves for each fibre. The total efficiency netic fields in a similar way. This increasespolarisation standard. Figure 5 shows thewith HARPSpol is somewhat lower due the signal-to-noise even further. The topcomparison of the HARPSpol polarisationThe Messenger 143 March 20119 10. Telescopes and Instrumentation Piskunov N. et al., HARPSpol The New Polarimetric Mode for HARPS Figure 7. Spectropolarimetry of a K2 dwarf planet- hosting star Eri taken with HARPSpol. Circular polarisation profiles (left) are marked with observa- tion times in days. Derived line-of-sight field strength and uncertainty in Gauss are shown against time (in Julian Day) on the right.spectra of this star with those takenAnother example is a chromosphericallya pipeline producing science-grade datawith the ESPADONS spectropolarimeter active cool dwarf Eri. This nearby star products. The tests and applications(Donati et al., 2006) at the Canadaharbours at least two planets and a dustto various types of objects have demon-France Hawaii Telescope (CFHT) with abelt in orbit around it. Polarisation meas- strated high sensitivity and a low levelresolving power of 67 000. urements of stars hosting planets may of systematic effects, making HARPSpol provide an important check for the pres-an ideal tool for detecting and studying ence of starspots that can mimic radial weak magnetic fields, reconstructing fieldHARPSpol: First resultsvelocity variations. Detection of polarisa- topology and many other magnetic phe- tion can reveal signatures of starplanet nomena.One of the obvious applications of magnetic interactions. Our polarisationHARPSpol is in the study of the topology measurements for Eri are presentedof magnetic fields on chemically pecu- in Figure 7. Again, we applied the LSDReferencesliar (CP) stars. The goal is to understand technique to enhance the signal-to-noiseBagnulo, S. et al. 2009, PASP, 121, 993the relationship between the field geome-ratio and we see an unambiguous sig-Donati, J.-F. et al. 1997, MNRAS, 291, 658try and the surface/depth distribution nal in circular polarisation. A simplisticDonati, J.-F. et al. 2006, Solar Polarization 4,of chemical elements. This task requires interpretation with a longitudinal fieldASP Conf. Series, 358, 362 Kochukhov, O. et al. 2010, A&A, 524, 5a series of observations well spread geometry shows field strength changingKotov, V. A. et al. 1998, ApJ, 116, 103over the rotation period so as to see allfrom 5.8 to +4.7 Gauss with medianMayor, M. et al. 2003, The Messenger, 114, 20visible parts of the stellar surface. Figure uncertainty of 0.1 Gauss! These valuesRamelli, R. et al. 2010, SPIE, 7735, 16 shows an example of one measure- are comparable to the disc-averaged Semel, M. et al. 1993, A&A, 278, 231 Snik, F. et al. 2008, SPIE, 7014, 22ment in such a series for a cool magneticmagnetic field of the Sun (Kotov et al.,Snik, F. et al. 2010, arXiv: 1010.0397CP star HD 24712. Circular and linear1998).polarisation were detected in all 13 phasescovering the whole stellar rotation (badweather prevented the collection of oneProspectsset of circular polarisation data) and onecan easily follow the evolution of polari- HARPSpol adds powerful polarimetricsation spectra with stellar rotation. Thecapabilities to the suite of ESO high-low level of the noise makes the dataresolution spectroscopic instruments. Itquite adequate for reconstructing the fieldis fully integrated into the ESO opera-topology.tional environment and is equipped with10 The Messenger 143 March 2011 11. Telescopes and InstrumentationTests of Radiometric Phase Correction with ALMABojan Nikolic1 than temperature fluctuations. ALMA is observed astronomical data can be cor-John Richer1 attempting to correct the effects of these rected for the effect of path fluctuations.Rosie Bolton1fluctuations through a combination ofRichard Hills 2two techniques: frequent observations of calibration sources; and direct measure- Water vapour radiometers ment of atmospheric properties along1Astrophysics Group, Cavendishthe line of sight of each of the 54 12-metre The water vapour radiometers (WVRs)Laboratory, University of Cambridge, diameter telescopes using mm-waveare the devices that measure accu-United Kingdom radiometers that measure emission of the rately the absolute brightness of down-2Joint ALMA Observatory, Santiago,183GHz water vapour line. ALMA is welling radiation along the lines of sight ofChilethe first telescope to employ phase cor- the antennas. The prototype WVRs for rection based on mm-wave water vapourALMA were developed by a collaboration radiometers. between the University of CambridgeOf the many challenges facing ALMA, and Onsala Space Observatory. Afterone of the greatest is overcomingWater in the atmosphere is poorly mixedsuccessful laboratory and field testing ofthe natural seeing limit set by the atmos- and the concentration (and phase) of the prototypes, an industrial partnerphere to achieve very high resolutionwater varies rapidly with position in the(Omnisys Instruments AB, Sweden) wasimages. Its longest antenna separationsatmosphere and with time. The underly- contracted for delivery of the produc-(baselines) permit ALMA to synthesiseing reason for this is of course that alltion units. The production stage is nowthe effect of a single antenna with athree phases of water are accessible inalready fully complete and ALMA hasdiameter exceeding 15km, but an ac the range of temperatures and pres-taken delivery of radiometers for all of thecurate radio adaptive optics systemsures typical on the ground and in the planned 54 12-metre antennas.is required to ensure ALMAs imagesatmosphere, leading to various localisedare diffraction limited. With initial test sources and sinks of water vapour. The ALMA radiometers are uniquedata now available from the first ALMA Even at a very high and dry site likeamong the radiometers used for phaseantennas in Chile, we describe current ALMA, changes of up to 50 % in line-of-correction in that they measure skyprogress towards this goal.sight water vapour can be observed brightness around 183 GHz, as opposed in a matter of minutes. Additionally, waterto 22 GHz, which is the spectral region vapour has a high effective refractive where most other WVR systems areAtmospheric limitations to radio index at mm and sub-mm wavelengths:designed to observe. This has a numberastronomyone millimetre of precipitable water of advantages, primarily based on the vapour retards radiation by an equivalentvery high strength of the water vapourALMA aims to synthesise an antenna of about seven millimetres of path inline at 183 GHz (see Figure 1 for plots ofwith an effective diameter of over 15 km:vacuum. The combination of poor mixing brightness in typical conditions), whichthis would have a diffraction-limitedand high refractive index leads to a is about 150 times stronger than the lineresolution of 15 milliarcseconds at a fre- corruption of the wavefront of incomingat 22 GHz. This means that fluctuationsquency of 300 GHz. (Note, however, astronomical radiation. When observing in water vapour content produce muchthat for most projects with ALMA, we with an aperture synthesis array likehigher, more readily observed fluctuationsanticipate that a more modest resolution ALMA, these wavefront errors lead to in the observed brightness at this fre-of 50100 milliarcseconds will be re-phase errors in the recorded visibilities. quency. Besides this, the high strengthquested by scientists.) In comparison, theof the line means that radiation fromuncorrected radio seeing at this fre-In order to correct for these errors, each sources other than atmospheric waterquency would typically limit the resolutionof ALMAs 12-metre diameter antennas vapour has a smaller influence on theof images to 700 milliarcseconds if no has an accurate millimetre-wave radiome- predicted phase corrections. For exam-adaptive optics corrections were applied ter that measures the radiation pas- ple, clouds, spill-over past the primary(see Evans et al., 2003).sively emitted by water molecules in the reflector of the antenna and man-made atmosphere along the line of sight ofradio frequency interference (RFI) all haveThe seeing at sub-millimetre and milli-the antenna. The radiometers cover fre-a smaller effect relative to the strength ofmetre wavelengths arises due to atmos- quencies around the 313 -> 220 rotationthe line.pheric (specifically, tropospheric) insta- line of the para water molecule, which isbilities that lead to fluctuations of thecentred at 183.3 GHz. This line lies about Measurements at these higher frequen-refractive index and consequent path 200K above the ground state and so is cies do, however, also present a numbererrors in the propagating wavefront. Asideal for tracing atmospheric properties.of challenges:explained in a previous Messenger arti-The principle of radiometric phase cor-1. Design and production of the hardwarecle (Nikolic et al., 2008), the processrection is that these measurements canis more complex and expensive,is analogous to that affecting the optical be used to compute the quantity of waterrequiring custom components and highseeing, but the dominant contributionvapour along the line of sight of eachprecision machining.to the refractive index fluctuations is from antenna and, consequently, the equiva- 2. Calibration is more difficult as it needsinhomogeneities in water vapour, ratherlent path error. Using these estimates theto be based on very frequent (10 Hz The Messenger 143 March 2011 11 12. Telescopes and Instrumentation Nikolic B. et al., Tests of Radiometric Phase Correction with ALMAFigure 1. The watervapour line at 183 GHz.The upper left, upperright and lower left pan-els show how the simu-lated brightness of theatmospheric 183 GHzwater vapour line varieswith changes in totalcontents of the watervapour, the atmospherictemperature and atmos-pheric pressure. Thelower right panel showsthe nominal filter pass- bands for the ALMA183 GHz water vapourradiometers. The detec-tion system is double-sideband and so onlythe average signal of thetwo filters symmetricaround the line centre ismeasured. in the case of ALMA) observation of we read out at 1 Hz, which is fast enough and we do not need to try to retrieve the physical internal calibration loads.to capture essentially all the path varia-total extra path due to the water vapour3. The water vapour line is close to satu- tions.in the atmosphere. Additionally, frequent ration and thus subject to non-linear observation of point-like sources will effects, leading to significantly moreThe task of the phase correction soft-allow ALMA to calibrate the expected complex software requirements.ware is to turn these 1 Hz measurements zero phase and it is only the departures in four filters into phase rotations to befrom this that are important.Over the past twelve months, extensive applied to the observed astronomical sig-testing of the first WVR systems has nal. The first step in the analysis is to use The close relationship between fluctua-been carried out at the ALMA site. The the four observed sky brightness tem- tions in sky brightness and the pathpreliminary results of these tests suggest peratures, together with ancillary weathererrors is illustrated in Figure 2, which isthat the development and productioninformation, to make an inference about based on recent observations by ALMA.stage has successfully met these chal- the total quantity, temperature and pres- During this observation, the telescopelenges. So far, the units installed on the sure of the water vapour. This is impor-was tracking a quasar (i.e., a point-likeALMA antennas appear to be perform-tant because the profile of the water va- source) at a known location on the sky,ing well in terms of noise, stability andpour line is a strong function of these so for a perfect interferometer we wouldreliability. parameters, as shown in Figure 1, and expect to measure visibilities with con- because the near-saturation of the line stant phase and amplitude. The phase means that the observed sky brightnesswe actually measure is therefore an esti-Techniqueis not in general linearly related to the mate of the differential path along the total path error. two lines of sight due to atmosphericThe WVRs provide measurements of fluctuations. This is plotted on the hori-sky brightness in the four filters illustrated The second stage of analysis is to turn zontal axis of the diagrams, and on thein the lower right plot in Figure 1. Asthe fluctuations in the observed skyvertical axis we plot the difference inALMA WVRs employ a double-sideband brightness into estimates of fluctuation of observations by the WVRs on these twomixing system, only the average bright-effective path to each of the antennasantennas. What can be seen in Figure 2ness of the sky at frequencies symmetric in the array. We only consider the fluctua- is that there is a high degree of correla-around the centre of the line is meas- tions because, as an interferometer,tion between the two quantities, meaningured. The maximum readout frequencyALMA is sensitive to only the difference in that as long as we can forecast the slopefrom the WVRs is 5 Hz, although normally path errors to each of the antennas of this correlation then we can convert12The Messenger 143 March 2011 13. Figure 2. Correlation between the atmos- pheric path error esti- mated from observa- tions of bright point-like objects (horizontal axis) and the differenced WVR signal (vertical axis). Each plot is a two- dimensional histogram where the colour scale shows how many points fall in each bin. The four panels correspond to the four channels of the radiometers (14 desig- nated by the axis label). fluctuations of WVR outputs to path fluc- (This programme also supports develop- The software is easy to distribute as atuations for a general observation. ment of Band 5 receivers for ALMA [seebinary package that works in conjunc-Laing et al., 2010], and on-the-fly interfer- tion with CASA.The final stage of the analysis is to turnometry techniques).estimates of path fluctuations into a The software (wvrgcal) for phase cor-phase correction that needs to be applied The software we have been developing is rection is available freely under the Gnuto the astronomical data. This is gener-designed primarily for off-line phase cor-Public License in both source code andally straightforward, although it is impor- rection, i.e., it operates on the observedbinary formats1. We also operate atant to take into account the dispersivedata after these have been stored onmailing list2 for discussion, improvementeffects of the atmosphere and also to disk. Some of the principal features of our suggestions and community support ofensure that rotations applied using esti- software are: the software.mates derived from WVR data interact It is closely integrated with the officialcorrectly with other calibrations applied ALMA off-line data reduction suiteto the astronomical data. (CASA), which allows it to be used in a Tests of phase correctionstraightforward manner by scientists. The software has a rapid developmentSince about January 2010, ALMA hasDevelopment of phase correction soft- cycle, with new features and improved been collecting significant amountsware for ALMA under FP6 algorithms appearing regularly. of test observations designed to measure It uses a robust Bayesian statistical the effectiveness of WVR phase correc-Our recent involvement in WVR phase inference framework to derive optimal tion and to guide the further develop-correction for ALMA has been primar-corrections.ment of algorithms used to translate skyily through development of software and When certain WVRs are missing frombrightness measurements to the phasealgorithms that process the raw dataan observation, the software hasrotations. In order to fit with the numerousobserved by the WVRs and use these to the ability to interpolate available data other ALMA commissioning activities,calibrate and correct the astronomicalto provide phase correction estimates most of these observations were takendata. This work is separate from the base-at those antennas lacking accuratewith the antennas in relatively compactline ALMA software and has been fundedWVR measurements. configurations, i.e., most data are withas an ALMA enhancement by the Euro- baselines in range 30100 m, with somepean Union Framework Programme 6. data on baselines of up to 600 m. These The Messenger 143 March 2011 13 14. Telescopes and Instrumentation Nikolic B. et al., Tests of Radiometric Phase Correction with ALMAdata have already provided a good dem-Figure 3. Path fluctua-tion estimated fromonstration of effectiveness of phase cor-WVR data for two ob -rection on these relatively modest base-serving sessions. Timeslines. However, we know that the phaseare expressed in UT,correction will be most challenging onso the upper panel cor-responds to night-time,long baselines (up to 15 km in length forwhile the lower panelALMA); this is because the root structure to a time around mid-function of the atmosphere increases as day. Note that the verti-roughly the 0.6 power of baseline on typi-cal scale is differentbetween the two panels.cal ALMA baselines. Long baseline testdata are awaited to investigate the effec-tiveness of the technique when ALMA ismaking its highest resolution images.Two typical examples of path fluctuationscomputed from WVR observations areshown in Figure 3. For these plots wehave used data from three antennas,shown by different colours in these plots.Since these are absolute path estimatesfrom the WVRs, it is the differencesbetween the three traces that correspond to the phase rotations to be applied.For these observations, the antennaswere relatively close to each other andtherefore these differences are quitesmall. These plots illustrate very well thewide variety of conditions that are pre-sent at the ALMA site: total fluctuationsare different by about two orders of mag- Figure 4. Test observa-tion of a sub-mm brightnitude between the two observations.quasar on a roughly 650-It can be seen that the total (peak-to- metre baseline withpeak) fluctuations on the upper panel ofALMA. The red line is theFigure 3 are about 50 m on timescalesphase (in degrees) of theobserved (complex) visi-of about five minutes; this is significantlybility on this baseline less than 350m, the shortest wave-note that for a quasar (orlength at which ALMA will observe. On other point-like) source atthe lower panel of Figure 3, the fluctua- the tracking centre of theinterferometer we expecttions are greater than 3.5 mm, i.e, theya constant phase in time.are larger than the longest wavelength at The blue line is the visibil-which ALMA will initially observe.ity phase after correctionof the data based on theFigures 4 and 5 show two examples WVR signals and usingthe wvrgcal program.of WVR phase correction at work. In bothplots, the red trace represents the phaseof the recorded visibilities while observ-Figure 5. Like Figure 4,ing a quasar. In the absence of atmos-this is a test observationof a strong quasar, butpheric and instrumental phase errors we on a baseline of aroundwould expect this phase to be constant60m and during stablein time the variations actually observedweather. The red line isare due to the combination of atmos-again the uncorrectedobserved phase (inpheric effects and instrumental errors. degrees) of the visibility,The uncorrected phase in Figure 4 is vary-while the blue line is theing by more than 360 degrees, i.e., byphase after WVR-baseda full rotation, which means that in thesecorrection. Note thechange of vertical scaleconditions it would not be possible tobetween Figure 4 andmake any measurements on faint sources. this figure.The blue line shows the phase aftercorrection using our wvrgcal software.14 The Messenger 143 March 2011 15. 4405o000202040406J2000 DeclinationJ2000 Declination0608081010121214144405o001605 38 51 .0h m s50 .6 50 .4 50 .2 50 .0s s ss49 .8s49 .6s05 h38 m51s.0 50 s.6 50 s.4 50 s.2 50 s.049 s.8 49 s.6J2000 Right Ascension J2000 Right Ascension Figure 6. Un-deconvolved images of a quasar (which It can be seen that the fringes in the mapto use, so they can automatically remove is unresolved) with ALMA in a very heterogeneousmade from corrected data (right panel)the distorting effects of the atmosphere configuration: four of the antennas were very close together while the fifth was about 650 metres away.are much sharper and have much higher and allow them to focus on the novel sci- The heterogeneity leads to the rapid modulationcontrast compared to the map made ence in their ALMA datasets. Nonethe- in the northsouth direction, which corresponds to from uncorrected data. After deconvolu- less, ALMA already has made great pro- the long baseline. The image on the panel on thetion (and completing the baseline cover-gress towards this goal, and can already left was made with no phase correction while the image in the panel on the right has had WVR phaseage) this increase in sharpness directlyclaim to have an effective and ground- correction applied. It can be seen that the phasecorresponds to an increase in resolutionbreaking adaptive optics system. fluctuations, when uncorrected, lead to an almostand fidelity. complete wash-out of fringes on the long baseline.AcknowledgementsFuture challenges The results shown in these plots are of course the The phase errors can be seen to be result of the efforts of many tens, if not hundreds, of reduced by an order of magnitude, to a The data presented in this article repre- people who have been involved in ALMA over the level where meaningful averaging of thesent by far the most extensive tests of years. Phase correction tests require everything inthe ALMA system to be working perfectly, so a great data can be done.the capabilities of 183 GHz phase correc- deal of credit is due to all of those involved, fromtion ever attempted. They demonstrate the designers of the systems to those keeping the The example shown in Figure 5 is lessthat the technique should increase signifi- observatory running in Chile. The specific work extreme the uncorrected data havecantly the sensitivity of ALMA, by reduc- described here, including the analysis of test dataand development of the wvrgcal program has a phase root-mean-square deviation ing the decorrelation caused by phase been carried out by the Astrophysics Group at the of about 20 degrees. However, even inerrors, and increase the fidelity of ALMA Cavendish Laboratory, University of Cambridge, these much more stable conditions, images by ensuring visibility phases areas part of the ALMA Enhancement Programme, an application of WVR phase correctionmore accurately measured. In addition,enhancement to the baseline ALMA project. Thiswork is funded by the European Unions Sixth leads to much improved phase stability.they should improve the efficiency of Framework Programme. Also notable in this example is that ALMA operations, by permitting observa- variations in uncorrected phase at longertions to take place when atmospheric timescales are also very effectively re- instabilities cause rapid large amplitude References duced by WVR phase correction. phase fluctuations. Evans, N. et al. 2003, Site properties and stringency, ALMA Memo Series, 471, The ALMA Project As an illustration of the effect of WVR- However, much work remains to be done.Laing, R. et al. 2010, The Messenger, 141, 41 based phase correction on imaging, It is vital that ALMA can demonstrate Nikolic, B. et al. 2008, The Messenger, 131, 14 in Figure 6 we show an un-deconvolvedthat its phase correction strategy works (dirty) map of a point source with to specification in a wide range of atmos-Links ALMA in an unusual configuration withpheric conditions and on baselines all1 north-south baselines much longerthe way out to the maximum allowed by Source code for wvrgcal available at: http://www.mrao.cam.ac.uk/~bn204/alma/wvrsoft.html than the others. We have made the mapthe configuration designs. In addition, 2Mailing list for wvrgcal updates: https://lists.cam. both with the raw data, and with the it remains a challenge to ensure that the ac.uk/mailman/listinfo/mrao-wvrgcal data after WVR-based phase correction. software tools are easy for astronomersThe Messenger 143 March 2011 15 16. Telescopes and InstrumentationGRAVITY: Observing the Universe in MotionFrank Eisenhauer1Frdric Chapron 2, 10 GRAVITY is the second generation VeryGuy Perrin 2, 10 Udo Neumann 3Large Telescope Interferometer instru-Wolfgang Brandner 3Leander Mehrgan 9ment for precision narrow-angle as-Christian Straubmeier 4Oliver Hans1 trometry and interferometric imaging.Karine Perraut 5 Grard Rousset 2, 10 With its fibrefed integrated optics,Antnio Amorim 6 Jose Ramos 3 wavefront sensors, fringe tracker, beamMarkus Schller 9Marcos Suarez 9stabilisation and a novel metrologyStefan Gillessen1Reinhard Lederer1concept, GRAVITY will push the sensi-Pierre Kervella 2, 10Jean-Michel Reess 2, 10tivity and accuracy of astrometry andMyriam Benisty 3 Ralf-Rainer Rohloff 3interferometric imaging far beyond whatConstanza Araujo-Hauck 4 Pierre Haguenauer 9is offered today. Providing precisionLaurent Jocou 5Hendrik Bartko1astrometry of order 10 microarcseconds,Jorge Lima 6 Arnaud Sevin 2, 10 and imaging with 4-milliarcsecondGerd Jakob 9 Karl Wagner 3resolution, GRAVITY will revolutioniseMarcus Haug1 Jean-Louis Lizon 9 dynamical measurements of celestialYann Clnet 2, 10Sebastian Rabien1objects: it will probe physics close toThomas Henning 3 Claude Collin 2, 10the event horizon of the Galactic CentreAndreas Eckart 4 Gert Finger 9black hole; unambiguously detect andJean-Philippe Berger 5, 9Richard Davies1measure the masses of black holesPaulo Garcia 6 Daniel Rouan 2, 10 in massive star clusters throughout theRoberto Abuter 9 Markus Wittkowski 9Milky Way; uncover the details of massStefan Kellner1Katie Dodds-Eden1accretion and jets in young stellarThibaut Paumard 2, 10Denis Ziegler 2, 10objects and active galactic nuclei; andStefan Hippler 3 Frdric Cassaing 7, 10probe the motion of binary stars, exo-Sebastian Fischer 4Henri Bonnet 9 planets and young stellar discs. TheThibaut Moulin 5 Mark Casali 9instrument capabilities of GRAVITY areJaime Villate 6Reinhard Genzel1 outlined and the science opportunitiesGerardo Avila 9Pierre Lena 2that will open up are summarised.Alexander Grter1Sylvestre Lacour 2, 10Armin Huber 31Max-Planck Institute for Extraterrestrial Fundamental measurements over a wideMichael Wiest 4 Physics, Garching, Germanyrange of fields in astrophysicsAxelle Nolot 5 2LESIA, Observatoire de Paris, CNRS,Pedro Carvas 6UPMC, Universit Paris Diderot, Much as long-baseline radio interfer-Reinhold Dorn 9 Meudon, Franceometry has tone, GRAVITY infrared (IR)Oliver Pfuhl13Max-Planck Institute for Astronomy, astrometry, with an accuracy of orderEric Gendron 2, 10Heidelberg, Germany 10 microarcseconds and phase-referencedSarah Kendrew 34Physikalisches Institut, University ofimaging with 4-milliarcsecond resolution,Senol Yazici 4Cologne, Germanywill bring a number of key advancesSonia Anton 6, 8 5UJFGrenoble 1/CNRS-INSU, Institut(Eisenhauer et al., 2008). GRAVITY willYves Jung 9 de Plantologie et dAstrophysique de carry out the ultimate empirical test toMarkus Thiel1 Grenoble, Franceshow whether or not the Galactic Centrelodie Choquet 2, 10 6Laboratrio de Sistemas, Instrumen- harbours a black hole (BH) of four millionRalf Klein 3tao e Modelao em Cincias e solar masses and will finally decide ifPaula Teixeira 6, 9 Tecnologias do Ambiente e do Espao the near-infrared flares from Sgr A* origi-Philippe Gitton 9 (SIM), Lisbon and Porto, Portugal nate from individual hot spots close toDavid Moch17ONERA, Optics Department (DOTA),the last stable orbit, from statistical fluc-Frdric Vincent 2, 10Chtillon, France tuations in the inner accretion zone orNatalia Kudryavtseva 3 8Centro de Investigao em Cinciasfrom a jet. If the current hot-spot interpre-Stefan Strbele 9 Geo-Espaciais, Porto, Portugaltation of the near-infrared (NIR) flaresEckhard Sturm1 9ESO is correct, GRAVITY has the potential toPierre Fdou 2, 10 10Groupement dIntrt Scientifique directly determine the spacetime metricRainer Lenzen 3 PHASE (Partenariat Haute rsolution around this BH. GRAVITY may evenPaul Jolley 9 Angulaire Sol Espace) between be able to test the theory of general rela-Clemens Kister1 ONERA, Observatoire de Paris, CNRStivity in the presently unexplored strongVincent Lapeyrre 2, 10 and Universit Paris Diderotfield limit. GRAVITY will also be able toVianak Naranjo 3unambiguously detect intermediate massChristian Lucuix 9BHs, if they exist. It will dynamicallyReiner Hofmann1 measure the masses of supermassive16The Messenger 143 March 2011 17. BHs (SMBHs) in many active galac-tic nuclei, and probe the physics of theirmass accretion, outflow and jets withunprecedented resolution. Furthermore,GRAVITY will explore young stellar obj-ects, their circumstellar discs and jets,and measure the properties of binarystars and exoplanet systems. In short,GRAVITY will enable dynamical measure-ments in an unexplored regime, and itwill increase the range and number ofastronomical objects that can be studiedwith the Very Large Telescope Interfer-ometer (VLTI) substantially. An overviewof the key experiments that will becomepossible with GRAVITY is illustrated onthe Telescopes and Instrumentation sec-tion page (p. 6, lower panel).A unique combination with the VLTIThe VLTI is the largest array of 8-metre-class telescopes that explicitly includedinterferometry in its design and imple-mentation. No other array is equippedwith a comparable infrastructure. TheVLTI, with its four 8-metre Unit Tele-scopes (UTs) and a total collecting areaFigure 1. (Upper) GRAVITY at the VLT Interferometer.Figure 2. (Lower) Working principle of GRAVITY.GRAVITY combines the light from four UT or AT tele- The beam-combiner instrument (bottom right) isof 200 m2, is the only interferometerscopes, measuring the interferograms for six base-located in the VLTI laboratory. The IR wavefront sen-to allow direct imaging at high sensitivity lines simultaneously, with a maximum baseline ofsors (bottom left) are mounted on each of the fourand high image quality. The VLTI is also200 metres. The insets depict the GRAVITY beam- UTs. The laser metrology is launched from the beamthe only array of its class offering acombiner instrument (middle), which is located in combiner and is detected at each UT/AT (top mid-the VLTI laboratory, and one of the four GRAVITY IR dle).large (2-arcsecond) field of view and thiswavefront sensors (left) for each of the UTs.unique capability will, for the first time,be utilised, providing simultaneous inter-Phase referenceferometry of two objects. This capability Objectallows narrow-angle astrometry with Wavefront referencea precision of order 10 microarcseconds.A second new and unique element of2 FoV Telescope #1Metrology fringes on secondaryGRAVITY is the use of IR wavefront sen-sors to observe highly obscured objects Metrologysuffering high extinction. GRAVITY isTelescope #2 sensorMetrologyalso the only instrument providing phase- Starreferenced complex visibilities, which is Starlightseparator 2 FoVDelay linea major advantage for the model inde-pendence and fiducial quality of interfero-metric maps. The combination of VLTI Laser Acquisitionand GRAVITY will be the world-leadingLaser guiding cameraMACAOfacility for many years to come.DMTip-tiltBeam combiner instrumentpupil dOPDIR wavefrontcontrol IO Beam combiner spectrometersensor AAdaptive optics assisted interferometric Dimaging and astrometry FiberB coupler CGRAVITY provides high precision narrow- Phaseshifter Metrologyangle astrometry and phase-referencedLaser Polarisation controlinterferometric imaging in the astronomi- OPD controlcal K-band (2.2 m). It combines the lightThe Messenger 143 March 2011 17 18. Telescopes and InstrumentationEisenhauer F. et al., GRAVITY: Observing the Universe in Motionfrom four UTs or Auxiliary Telescopes combiners (Jocou et al., 2010) and thehighlights from the ongoing prototype(ATs), measuring the interferograms fromcoherently combined light is dispersed in development: the worlds first K-bandsix baselines simultaneously (see Figure 1).two spectrometers (Straubmeier et al.,(2.2 m) integrated optics beam combinerThe instrument has three main compo-2010). A low resolution spectrometer pro- for four telescopes, a high-speed photon-nents: the IR wavefront sensors (Clnet etvides internal phase- and group-delay counting IR detector, and a novel laseral., 2010); the beam-combiner instru- tracking (Choquet et al., 2010) on the ref- metrology concept.ment; and the laser metrology system erence star, and thus enables long expo-(Bartko et al., 2010). Figure 2 gives ansure times on the science target. Three GRAVITYs beam combiner is an inte-overview of the GRAVITY instrument. For spectral resolutions with up to R ~ 4000grated optics chip, the optical equiva-clarity, only two of the four telescopes are implemented in the science spec-lent of a microelectronic circuit, whichi.e. one out of six baselines are shown.trometer, and a Wollaston prism providescombines several functions in a singlebasic polarimetry.component. It combines the advantagesThe GRAVITY IR wavefront sensors will of compactness and stability, and pro-be mounted in the Coud rooms ofGRAVITY will measure the visibility ofvides outstanding visibility accuracies.the UTs and will command the existing the reference star and the science object Integrated optics is widely used in tele-Multiple Application Curvature Adaptive simultaneously for all spectral channels, communications up to 1.6 m, but doesOptics (MACAO) deformable mirrors.and the differential phase between thenot cover the astronomically interestingThe system can work on either of the twotwo objects. This information will be usedK-band. GRAVITY has thus launched itsbeams (on-axis or off-axis) behind thefor interferometric imaging exploring own development programme betweenPRIMA star separators. Any additional the complex visibilities, and for astrome-IPAG, LETI, and CIP to port the technol-tip/tilt from the beam relay down to thetry using the differential phase andogy to longer wavelengths. Following aVLTI laboratory will be corrected by agroup delay. All functions of the GRAVITY series of prototypes implementing indi-dedicated laser-guiding system. Low fre-beam-combiner instrument are imple- vidual functions, we now have the worldsquency drifts of the field and pupil will mented in a single cryostat for optimum first K-band integrated optics beam com-be corrected by GRAVITYs internalstability, cleanliness, and thermal back- biner for four telescopes in hand (shownacquisition and guiding camera (Amorimground suppression. The internal path in Figure 3).et al., 2010). The interplay of these sys-lengths of the VLTI and GRAVITY aretems will guarantee an unperturbed andmonitored using dedicated laser metrol- The second major breakthrough forseeing-corrected beam at the entrance ogy. The laser light is back-propagated GRAVITY is the recent success in theof the beam-combiner instrument infrom the beam combiner and covers development of high-speed IR photon-the VLTI laboratory. The interferometricthe full beam up to the telescope spidercounting detector arrays. All currentinstrument will work on the 2-arcsecond above the primary mirror. astronomical IR fringe trackers and wave-(for UTs) or 4-arcsecond (for ATs) VLTI front sensors suffer from the high read-field of view. Both the reference star andout noise of their detectors, which is tenthe science object have to lie within thisHighlights from the instrument develop- or more electrons per pixel at framefield of view. The light of the two objects mentrates of a few hundred Hz. The GRAVITYfrom the four telescopes is coupled (Pfuhldetectors overcome this noise barrieret al., 2010) into optical fibres for modal A detailed description of GRAVITYs by avalanche amplification of the photo-filtering, to compensate for the differential subsystems can be found elsewhere electrons inside the pixels. Last year,delay and to adjust the polarisation. The (Gillessen et al. [2010] and aboveSELEX-Galileo and ESO demonstratedfibres feed two integrated optics beamreferences). Instead, we present a fewfor the first time a readout-noise of lessFigure 3. A recentbreakthrough in inte-grated optics is shown(left) and an exampleimage from the newavalanche photodiodedetector arrays (right);both to be used inGRAVITY.18The Messenger 143 March 2011 19. Figure 4. Uncovering the true nature of Sgr A* flares (upper three panels); probing spacetime close to the black hole event horizon (lower left); and measuring its spin and inclination (two lower right panels). GRAVITY will easily distinguish between the three most plausible flare scenarios: a jet (left), an orbiting hot spot (middle) and statistical fluctuation in the accretion flow (right). The detailed shape of the photo-centre orbit is dominated by general relativis- tic effects (lower left, from Paumard et al. 2008), and GRAVITY will thus directly probe spacetime close to the event horizon. The combination of time-resolved astrometry (lower middle) and photometry (lower right, from Hamaus et al., 2008) will also allow the spin and inclination of the BH to be measured.4 60 20 Primary Secondary image image4021520Flux (mJy) 10R/Rs as 0 0 205 2 Newtonian 40 Observed orbitorbit0 4 60 0 50 100 1502004 20 2 4 60 40 20020 40 60 Time (min)R/Rsasthan three electrons with their prototype answer with GRAVITY: What is the naturetuations in the accretion flow (Figure 4).detector array (see image in Figure 3). of the flares in Sgr A*? What is the spinThe jet model seems natural from theBased on this success, ESO and SELEX of a BH? How can we resolve the Para- presence of jets in active galactic nuclei.Galileo are currently developing a next dox of Youth of the stars in its vicinity?The orbiting hot-spot model would be ageneration detector, which is tuned toEven tests of fundamental physics maynatural explanation for the observedGRAVITYs wavefront sensor and fringe come into reach with GRAVITY: Does the quasi-periodicity in the light curves oftracker. Another example of a major theory of general relativity hold in the flares and associated changes of the IRbreakthrough is in GRAVITYs laserstrong field around SMBHs? Do BHs really polarisation. However, the long-term lightmetrology. It is based on a novel concept,have no hair?curves are well described by a pure, redand traces the starlight through the ob- power-law noise, indicating that statisticalservatory, to allow the optical path to be fluctuations in the accretion flow aremeasured at any desired point of the pupilUncovering the true nature of the SgrA*responsible for the observed variability.up to the primary mirror. This conceptflares Time-resolved astrometric measurementsand its implementation have been demon-with GRAVITY will settle the debatestrated in three technical runs at the VLTI.The Galactic Centre BH is surprisingly (Eckart et al., 2010). Even without push-faint its average luminosity is only ing GRAVITY to its ultimate performance,about 10 8 of the Eddington luminosity, the observed distribution of flare posi-Science cases for GRAVITY emitted predominantly at radio to sub- tions and its periodic variation will distin-mm wavelengths. On top of this quasi-guish between these models.In the following sections the science steady component there is variable emis-cases for GRAVITY are briefly outlined, sion in the X-ray and IR bands. Somebeginning with the broad range of sci-of this variable emission comes as flares, Measuring spin and inclination of theence opportunities that have opened uptypically a few times per day, lasting for Galactic Centre black holeat the Galactic Centre of the Milky Way.about one to two hours, and reaching theThe Galactic Centre is by far the closest brightness of massive main-sequenceThe mass of the Galactic Centre BH isgalactic nucleus and the best studied stars. The three most plausible explana- well known from stellar orbits. If the cur-SMBH (Genzel et al., 2010). There are still tions for the origin of these flares are:rently favoured orbiting hot-spot modela number of fundamental open issues a jet with clumps of ejected material; hot is correct, GRAVITY will take the nextand just to name a few that we want tospots orbiting a BH; or statistical fluc-step and measure its spin and inclination.The Messenger 143 March 2011 19 20. Telescopes and Instrumentation Eisenhauer F. et al., GRAVITY: Observing the Universe in MotionThese measurements are more difficultearly-type main sequence stars. It issignificantly enlarge the number of starsbecause the astrometric signature from currently not understood how these stars with known eccentricities, and will dis-the spin is a factor few less than the or- have formed or moved so close to the tinguish between the formation scenariosbital motion and lensing effects. However, SMBH, because the tidal forces shouldunambiguously (see Figure 5).the combined signal from the periodichave prevented in situ formation, andlight curves and astrometry is muchbecause these stars are too young tostronger. Already the simple correlation have migrated so far within the timescaleTesting general relativity in the strong fieldbetween the observed position variationof classical relaxation. Precise orbit regimeand flux variability is giving the first in- measurements with GRAVITY offer asights into the source geometry. The nextroute to resolving this Paradox of Youth.The unprecedented astrometric accuracystep is a simultaneous fit to the observed In particular measurements of the orbitalof GRAVITY may even allow the theorymotion and light curve to quantify the eccentricities can distinguish between of general relativity to be tested in the (sounderlying model parameters (Figure 4).the various scenarios. The currently far) unexplored strong field aroundFinally, the periodic flux can be used favoured Hills scenario, in which massiveSMBHs. The observed orbit of a hot spotto trace the orbital phase to coherently binaries are scattered down to the BHon the last stable orbit will be domi-co-add measurements from multipleand one component is ejected in a three- nated by strong gravitational effects likeflares, such that higher order signaturesbody interaction with the BH, will leadgravitational lensing and redshift (Fig-can be directly identified.to predominantly high eccentricities. In ure 4). GRAVITY observations of the flar- contrast, the competing migration sce- ing BH will thus directly probe spacetime nario, in which the stars migrate from cir-in the immediate vicinity of the eventResolving the Paradox of Youth of thecumnuclear stellar discs, will result mainly horizon of the BH. The stellar orbits willGalactic Centre starsin low initial eccentricities. First results be notably affected by higher order gen- from adaptive optics observations slightly eral relativistic effects, for example theMost stars in the central light-month of favour the Hills scenario, but the sig-relativistic periastron shift and the Lensethe Galactic Centre are young, massive nificance is still marginal. GRAVITY willThirring precession of the orbital angularmomentum around the BH spin axis 10 000 (Figure 6). These effects will be strongest NACOS31for stars within the central light-week, S19 0.4S27S12which will be observed with GRAVITY in 1000S5S29its interferometric imaging mode. In the 0.2 S2 S14 S17most optimistic case, GRAVITY may even S6a (as/yr 2 )1 S4be able to test the so-called no-hairDec ( )100 GRAVITY0theorem (Will, 2008), which states that a2S39 S21BH is fully characterised by its mass 0.2S18and spin. In particular the BH spin and its S13 10S8 S1quadrupole moment should be strictly 0.4 S9 S24 S33related. Since spin and quadrupole10moment couple differently to the inclina- 1tion of stellar orbits, they can be meas- 0 24 6 810 0.40.20 0.2 0.4 Years of observationRA ( )ured independently (see Figure6). 1.0Active galactic nuclei 0.8The standard unified model for activeCumulative Probabilitygalactic nuclei postulates that an ac-creting SMBH is surrounded by an ob- 0.6 Migrationscuring torus, whose orientation deter-Thermal mines if the central engine is hidden 0.4Figure 5. Solving the Paradox of Youth of the Galac-tic Centre stars. GRAVITY will be able to measureaccelerations, i.e. individual orbits, out to about 0.210 arcseconds distance from the SMBH (upper left)Hills and will significantly enlarge the number of S-stars(upper right) with precise eccentricities (from 0.0Gillessen et al., 2009). The improved eccentricitydistribution (lower) can distinguish between the vari-0.00.20.4 0.60.81.0 ous formation scenarios proposed for stars in theEccentricitycentral light-month.20 The Messenger 143 March 2011 21. a/R s Figure 6. Testing the theory of general1 10 100 100010 000relativity with stellar orbits. GRAVITY 100 00015 will observe the orbits of stars withinLense-Thirring(frame-dragging) the central light-week of the Galactic Centre by means of interferometric 10imaging (upper panels: dirty beam 1000 1000 with a resolution of four milliarcsec-Precession onds (left), simulated dirty imagey (milli-arcseconds)5(middle), cleaned image (right, from Paumard et al. 2008)). Stellar orbits t orb (yrs)t prec/t orb (illustrated in the lower left panel) 10 10 will be affected by the general relativ-0 istic periastron shift (red arrows) and the LenseThirring precession of the orbital angular momentum (blue 5 0.10.1arrows). For small distances to the BH, the timescale of these relativistica = 0.5 effects are short enough (lower right) 10M = 4 10 6 Mt orb to be in reach of GRAVITY (blue shaded area). 0.0010.001 15 10 50 5 10 0.01 0.1110 100 x (milli-arcseconds) a (milli-arcseconds)from the observers view or not. The Figure 7. A sketch of the prototypical activedirect proof that this absorber is really a0.5 pc/7 mas galactic nucleus oftorus, rather than another structure, is NGC 1068 (from Rabanstill pending. Indeed most resolved gase-et al., 2009). The gase-ous structures on the putative scale ous structures andof the torus appear more disc-like, for Maser disc dust emission on the scale of the putativeexample the maser disc, the radio contin-torus appear disc-like,uum emission and the mid-IR emission while the unified modelof the prototypical active galactic nuclei suggests a geometri- cally thick torus. Ob -NGC1068 (see Figure 7). Observing serving at NIR wave-six baselines simultaneously, GRAVITYlengths, GRAVITY willwill image the inner edge of the torus with Radioimage the inner edge ofunprecedented quality, where the dust contiuum the absorber, putting strong constraints onis close to the sublimation limit. GRAVITY the absorber geometry.will thus put strong constraints on theabsorber models. These models are very Mid-infraredmuch inspired by the observations ofNGC 1068, but the few active galacticcontinuumnuclei with interferometric observationsshow a puzzling variance. GRAVITY willsignificantly extend the sample to finallydraw statistically sound conclusions. The Messenger 143 March 2011 21 22. Telescopes and InstrumentationEisenhauer F. et al., GRAVITY: Observing the Universe in Motion100 000 Typical angular acceleration (as/yr 2 ) 10 0001000GalacticCentre 100GRAVITY, 2 year programme OmegaCentauri101 GRAVITY, 10 year programme NGC 62660.1The Arches NGC 6388M15 0.010.01 0.11Angular distance from central black hole ( ) Figure 8. Discovering IMBHs in star clusters. The left panel shows the globular cluster Cen and a zoom into its centre (from Anderson et al., 2010). The blue circle has a radius of 1 arcsecond. The statistical analysis of the velocity dispersion is limited by hav- ing only a few stars within the sphere of influence of the BH. GRAVITY will make a clear-cut case in a few suitable clusters by measuring the accelerations of individual stars (right), directly probing the central gravitational potential.Even the broad-line region may come inmass BHs in massive, dense star clus-nearby systems, for which a suitablereach for GRAVITY. It is seen in thoseters. Recent searches in globular clusters astrometric reference star is available.active galactic nuclei for which we have ashow evidence for such IMBHs (Fig- Combined with spectroscopy, thesedirect view onto the SMBH. The size ure 8). However, the sphere of influence observations will provide the orbital ele-of the broad-line region can currently only of the postulated BHs is typically lessments and distance of the system, asbe measured indirectly, looking at thethan a few arcseconds, such that onlywell as the mass of the two components.time delay between the variations of thea few stars are available for these statisti-In addition GRAVITY will characteriseultraviolet continuum and the emissioncal studies. GRAVITY will dramatically the wind from the stellar companion at alines. Broad-line regions of nearby activechange this situation in a few suitablescale of a few stellar radii. The physicalgalactic nuclei are typically smaller thancases for which accelerations can be de- properties of this wind are particularly0.1 milliarcseconds, and thus too small totected, thus directly probing the gravi- interesting as it is the main source forbe resolved in GRAVITYs images. Buttational potential without suffering fromfeeding the compact object.the astrometric accuracy of GRAVITY willthe small number statistics of velocityallow measurement of the velocity gra-dispersion measurements.dient across it. This will strongly constrainMasses of the most massive stars andthe broad-line region geometry and brown dwarfsdetermine dynamically the mass of the X-ray binariescentral BH.There is still a discrepancy of up to a fac-X-ray binaries are the best place to study tor of two in the mass estimates for theneutron stars and BHs. These neutron most massive main sequence stars. It isIntermediate mass black holes stars and BHs are very faint when iso- not known what the maximum allowedlated, but they can be observed as partmass for a star is. Comparison of spectraThe tight correlation between the bulge of a X-ray binary, some of which are with atmospheric models yields uppermass of a galaxy and the mass of thebright enough for GRAVITY. We expect mass limits of typically 60 MA, whereascentral SMBH suggests that the rapidthat it will be possible to detect the orbital evolutionary tracks and observed lumi-formation of a spheroidal stellar systemdisplacement from the compact compan-nosities suggest a mass of up to 120 MA.also collects up to about 1 % of the initialion in the interferometric closure phase.Clearly, dynamical mass estimates aremass in a central BH. Such a core col-Even the absolute astrometric displace-required. Quite a number of spectro-lapse and collisional build-up may have ment of the binary systems photo-centre scopic binary O-stars are known in thealso led to the formation of intermediate will be observable with GRAVITY in a few cores of starburst clusters like Arches,22The Messenger 143 March 2011 23. 30 Doradus and the Galactic Centre.mass regime. GRAVITY will probe many understanding of jet formation. The rele-GRAVITY will resolve some of the longermore multiple systems like AB Dor C, de- vant processes take place withinperiod spectroscopic binaries, and willriving the individual component masses,about one astronomical unit from themonitor the astrometric motions of the and even probe the sub-stellar compan- star, which at the typical distance to thephoto-centres for the short period, closeions themselves for binarity, thus clarify-nearest star-forming regions of aboutbinaries. In this way, GRAVITY will di-ing this situation.150 pc translates into an angular size ofrectly yield dynamical mass estimates for about 6 milliarcseconds, slightly largermany of these systems, and finally pro- than GRAVITYs 4-milliarcsecond angularvide the crucial input required to calibrate Jet formation in young stars resolution. By repeatedly imaging thestellar evolutionary tracks.time-dependent ejection just outside the Jets are omnipresent in the Universe,engine at high spectral resolution,The situation is similar for brown dwarfs, from gamma-ray bursts to active galac- GRAVITY will provide key observationalwhich are the lowest mass stars. Mosttic nuclei, from young stars to micro- tests of time-dependent jet simulationscurrent mass estimates are based onquasars. Understanding the formation of(Figure 9). Furthermore, the astrometricevolutionary models and model atmos- jets is still one of the big open chall- signal across the emission line willpheres, which have not yet been accu-enges of modern astrophysics. It is nowdirectly probe the central engine on therately calibrated by observations. Dynam-known that jets are powered by mag-sub-milliarcsecond scale, i.e. well withinical masses for brown dwarfs have only neto-hydrodynamic engines, tapping the one astronomical unit.been derived for a few objects. In gen-energy of the accretion disc. Youngeral, the observed masses for sub-stellarstars are ideal objects to study these pro-objects with ages older than a few cesses at the highest resolution. Matter Planet formation in circumstellar discs100million years seem to be in good from the disc surface couples to theagreement with theoretical models. But open, highly inclined stardisc magnetic Circumstellar discs are the cradles ofthere are significant uncertainties andfield lines, and is accelerated up toplanet formation. Planets are thoughtdiscrepancies for the very young, very the Alfvn surface. The rotating magneticto form rapidly in a few million yearslow mass objects like AB Dor C. If indeedfield lines then become more and morethrough the fast evolution of the discthese objects are more massive thantwisted, wind up and collimate the jet.structure. Dust processing, settling andindicated by stellar evolutionary models,But surprisingly, some stellar jets arecoalescence are accompanied by anmany putative planets would be ratherfound only on one side of the disc. Clearly, increase of particle size, leading to thein the brown dwarf than the planetarysome basic ingredient is missing in ourformation of planetesimals that eventuallyaggregate to form planetary systems.The planet formation process is expectedto leave strong imprints in the disc struc-ture, such as inner disc clearing, gapopening and tidally induced spiral struc-tures. GRAVITY will hunt down all thesesigns. Its unique sensitivity in the NIRwill allow the sample of observed youngstars to be increased towards the poorlyexplored solar-mass regime. This willbe done at sub-astronomical unit resolu-tions for the closest star-forming regions.It will reveal the disc structure evolution,the so-called transitional step, search fordisc disruption signatures, and be usedto probe the presence of hot, young sub-stellar and planetary companions.Astrometric planet detectionMany hundreds of exoplanets have beendetected to date, mostly from radialFigure 9. 3D simulation of a large-scale jet from anearby young star. In this picture the bow shock haspropagated roughly 400 milliarcseconds out fromthe jet engine (from Staff et al., 2010).The Messenger 143 March 2011 23 24. Telescopes and Instrumentation Eisenhauer F. et al., GRAVITY: Observing the Universe in Motion10.000 Figure 10. The realm of GRAVITY exoplanet search among very low-mass stars is shown. The1000.0 blue shaded area depicts the discovery space of Direct imagingGRAVITY for planets around a late M dwarf at a distance of 6 pc. The green and red areas indicate roughly the parameter space probed by radial veloc- Minimum detectable mass (MEarth ) 1.000Minimum detectable mass (MJup ) ity observations and direct imaging.100.0 GRAVITY 0.100Radial velocity 10.0measurements1.0 0.0100.1 0.001 0.01 0.101.00 10.00100.00 Separation a (AU) velocity measurements and photometricTransiting exoplanetsOn sky in 2014 transit observations. However, these methods are biased towards detecting The transit of a planet in front of its host The GRAVITY project emerged from massive planets in close orbits. More- star causes an apparent motion of theESOs second generation VLTI instrument over, radial velocity measurements alone photo-centre of the star and introducesworkshop in 2005. Following the initial cannot provide the inclination of thea slight asymmetry in the image of the Phase A study in 2006/7, ESOs Science orbit, and can thus only give a lower limitstar. The former effect can be measuredand Technical Committees recommen- for the mass of the planet. In contrast, using GRAVITYs astrometric observingdation and ESO Councils endorsement the reflex motion of a star observed bymode, the latter effect can be seen in the in 2008, and the preliminary design astrometry allows the orbital solution toclosure phases of the interferograms review in 2010, the project is currently in be retrieved, resulting in an unambiguous(van Belle, 2006). GRAVITY observationsits final design phase. First astronomical measurement of the mass of the planet. of such transits have the potential to light at the VLTI is planned for 2014.measure the radius of the planet and its Astrometric planet detection is also a sci-parent star. entific goal of the PRIMA facility, cur-References rently being commissioned at the VLTIFor a star like HD 189733, a 0.8 MA star Amorim, A. et al. 2010, Proc. SPIE, 7734, 773415 (Delplancke, 2009). While planet searchesat a distance of about 20 pc, and itsAnderson, J. et al. 2010, ApJ, 710, 1032 with PRIMA mostly target isolated starsJupiter-sized planet on a very close, two- Bartko, H. et al. 2010, Proc. SPIE, 7734, 773421 or wide binaries, GRAVITY will focus day orbit, the apparent motion is aboutvan Belle, G. 2008, PASP, 120, 617 Choquet, E. et al. 2010, Proc. SPIE, 7734, 77341Z on detecting brown dwarfs and exoplan- 10 microarceconds. This is at the limit of Clnet, Y. et al. 2010, Proc. SPIE, 7736, 77364A ets in close binary systems. For Sun-likeGRAVITYs capability, but transiting plan- Delplancke, F. 2008, NewAR, 52, 199 stars, GRAVITYs survey volume would ets around later-type dwarfs would beEckart, A. et al. 2010, Proc. SPIE, 7734, 77340X extend out to more than 200 pc. Even the easier to detect. This type of measure-Eisenhauer, F. et al. 2008, The Power of Optical/IR nterferometry:RecentScientificResultsand2ndI much fainter M-stars, with just 20 % ment will also give the position angle of Generation Instrumentation, ed. Richichi, A. et al., of the mass of the Sun, can be observedthe orbit on the sky, which, combined ESO Astrophysics Symposia, 41, 431 out to about 25 pc (Figure 10). GRAVITYwith the direction and amount of polarisa- Genzel, R. et al. 2010, RvMP, 82, 3121 has the potential to detect exoplanets tion of the light reflected by the planet, Gillessen, S. et al. 2010, Proc. SPIE, 7734, 77340Y Gillessen, S. et al. 2009, ApJ, 692, 1075 as small as three Earth masses aroundmight ultimately even place constraintsJocou, L. et al. 2010, Proc. SPIE, 7734, 773430 an M5V star at a distance of 5 pc, oron the distribution of surface features like Paumard, T. et al. 2008, The Power of Optical/IR less than two Neptune masses around an clouds and weather zones. nterferometry:RecentScientificResultsand2ndI M3V star at a distance 25 pc.Generation Instrumentation, ed. Richichi, A. et al.,ESO Astrophysics Symposia, 41, 431 Pfuhl, O. et al. 2010, Proc. SPIE, 7734, 77342A Raban, S. et al. 2009, MNRAS, 394, 1325 Staff, J. E. et al. 2010, ApJS, 722, 1325 Straubmeier, C. et al. 2010, Proc. SPIE, 7734, 773432 Will, C. M. 2008, ApJL, 674, 25 24 The Messenger 143 March 2011 25. Telescopes and InstrumentationThe E-ELT has Successfully Passed the Phase BFinal Design ReviewRoberto Gilmozzi1MacMynowski (Caltech/TMT), BuddyIn the meantime, Brazil has signed theMarkus Kissler-Patig1Martin (Steward Observatory Mirror Lab),formal accession agreement becom- Harald Nicklas (University of Gottingen), ing ESOs 15th Member State (see the Roberto Ragazzoni (INAF), Francoisannouncement by Tim de Zeeuw on1ESORigaut (Gemini), Luc Simard (HIA/TMT),p. 5 and the ESO press release 10502), Doug Simons (Gemini) and Larry Steppbringing new resources and skills to the (TMT). Six weeks before the review, organisation at the right time for themThe European Extremely Large Telescope the Telescope Construction Proposal to make a major contribution to this excit-(E-ELT) recently achieved a critical mile- with over 300 ancillary documents was ing project.stone by passing its Phase B final designmade available to the Board, to whichreview. The top-level question posed toit responded with over 300 written ques-All the pieces are now in place to go forthe Review Board was whether the tech- tions and comments. The review ranCouncil approval this year and to breaknical maturity of the design of the E-ELTover four days during which the designground on Cerro Armazones in 2012,is sufficient to warrant the programme was probed in great depth. The outcomethen head for first light of the biggest eyeentering the construction phase. It waswas praise, constructive feedback and on the sky at the end of this decade.the unanimous conclusion of the Review the unanimous agreement that the E-ELTBoard that the answer to this question isis ready to enter the construction phase.Yes. Links The Review Board n