Professional Documents
Culture Documents
The Messenger 178
The Messenger 178
AdOptica/ESO
Michele Cirasuolo 1 Thanks to the combined use of M4 and
Marc Cayrel 1 M5, the optical system is capable of
Roberto Tamai 1 correcting for atmospheric turbulence
Aglae Kellerer 1 and the vibration of the telescope struc-
Lorenzo Pettazzi 1 ture itself induced by motion and wind.
Paul Lilley 1
Pablo Zuluaga 1 This adaptive capability is crucial to
Carlos Diaz Cano 1 allowing the ELT to reach its diffraction
Bertrand Koehler 1 limit, which is ~ 8 milliarcseconds (mas) in
Fabio Biancat Marchet 1 the J-band (at λ ~ 1.2 μm) and ~ 14 mas
Juan Carlos Gonzalez 1 in the K-band. In so doing the ELT will
Mauro Tuti 1 be able to yield images 15 times sharper
+ the ELT Team than the Hubble Space Telescope and
with much greater sensitivity. Translated
into astrophysical terms this means
1
ESO opening up new discovery spaces, from Figure 1. Rendering of the M4 adaptive mirror unit
for the ELT.
exoplanets closer to their stars, to black
holes, to the building blocks of galaxies
The Extremely Large Telescope (ELT) is both in the local Universe and billions of consortium name of AdOptica. Many
at the core of ESO’s vision to deliver the light years away. For example, the ELT 8-metre telescopes now have a metre-
largest optical and infrared telescope will be able to detect and characterise scale adaptive mirror. The same tech
in the world. Continuing our series of extrasolar planets in the habitable zone nology is now being adapted to serve the
Messenger articles describing the opti- around our closest star Proxima Centauri, ELT, in order to produce a mirror with an
cal elements of the ELT, we focus here or to resolve giant molecular clouds (the area five times larger. The M4 mirror uses
on the quaternary mirror (M4), a true building blocks of star formation) down to the same principle as a loudspeaker; the
technological wonder; it is the largest ~ 50 parsecs in distant galaxies at z ~ 2 mirror is made of a very thin shell levitating
deformable mirror ever made. In combi- (and even smaller structures for sources 100 microns away from its reference sur-
nation with M5, M4 is vital to delivering that are gravitationally lensed by fore- face (this corresponds to the thickness
the sharp (diffraction-limited) images ground clusters) with an unprecedented of a standard A4 sheet of paper) and it
needed for science by correcting for sensitivity. acts like a membrane which deforms
atmospheric turbulence and the vibra- under the effect of about 5000 voice coil
tions of the telescope itself. Here we actuators. A voice coil actuator is a type
describe the main characteristics of M4, The quaternary mirror (M4) of direct drive linear motor and the name
the challenges and complexity involved “voice coil” comes from one of its first
in the production of this unique adaptive M4 is the main adaptive mirror of the tele- historical applications, vibrating the paper
mirror, and its manufacturing status. scope. The term “adaptive mirror” means cone of a loudspeaker. It consists of a
that its surface can be deformed to cor- permanent magnetic field assembly and
rect for atmospheric turbulence, as well a coil assembly. The current flowing
Background: how the ELT works as for the fast vibration of the telescope through the coil assembly interacts with
structure induced by its motion and the the permanent magnetic field and gener-
Let’s briefly recall how the ELT works. wind. In the case of M4, more than 5000 ates a force that can be reversed by
The optical design of the ELT is based on actuators are used to change the shape changing the polarity of the current.
a novel five-mirror scheme capable of of the mirror up to 1000 times per second.
collecting and focusing the light from Depending on the current injected into
astronomical sources and feeding state- In combination with the M5 mirror, M4 the coil the mirror can be pushed or
of-the-art instruments for the purposes of forms the core of the adaptive optics of pulled up to a distance of 90 microns
imaging and spectroscopy. The light is the ELT. With a diameter of 2.4 metres, from its mean position. With the help of
collected by the giant primary mirror M4 will be the largest adaptive mirror ever a very fast and precise set of capacitive
39 metres in diameter, relayed via the M2 built. By comparison, current adaptive sensors and amplifiers that are co-located
and M3 mirrors (each of which has a mirrors are just over 1 metre in diameter, with the voice coil actuators, the mirror’s
diameter of ~ 4 metres) to the M4 and M5 for example the 1.1-m M2 adaptive sec- position is measured 70 000 times per
mirrors that form the core of the adaptive ondary on the VLT UT4 telescope (Yepun). second to an accuracy of a few tens of
optics of the telescope; the light then nanometres (the size of the smallest virus)
reaches the instruments on one or other Adaptive mirror technology was trans- with the actuators being driven up to
of the two Nasmyth platforms. This lated into an industrial product for astron- 1000 times per second.
design provides an unvignetted field of omy more than two decades ago by
view (FoV) of 10 arcminutes in diameter the Italian companies Microgate s.r.l and M4 is made of several state-of-the-art
on the sky, ~ 80 square arcminutes ADS, internationally known under the components, the mirror and its reference
structure being two of the most critical The back of the reference structure is
Mersen Boostec
ones. The mirror is an assembly of six supported by a 12-point whiffletree and
optically polished thin shells, or petals, laterally at six points on the mirror edge.
made of the low-expansion glass-ceramic The overall M4 sub-system is mounted
Zerodur© (manufactured by Schott on six position actuators (a hexapod sys-
GmbH). The six petals are obtained from tem), which provide the fine alignment
a 35 mm-thick blank, which is polished of the mirror. It is further mounted on a
and thinned down to a thickness of less rotating mechanism (called a switcher)
than 2 mm — necessary to achieve the which is used to select the Nasmyth
desired flexibility for shaping the mirror — focus to which the light will be directed.
and then finally cut into a precise shape
by Safran Reosc (France; see Figure 2).
Manufacturing the M4
In order to adjust the shapes of the thin
shells, a rigid and sufficiently accurate Safran Reosc (France) started to manu-
flat reference structure is also needed to facture the thin segment mirrors in 2017
hold the petals. This structure must be and four thin shells are now ready for Figure 4. Detail of the M4 reference body.
stiff enough to provide a good reference integration in Italy. The remaining eight
surface, whatever the orientation of the shells still need to be delivered in order
telescope. It also needs to hold all the to have two sets of six shells each (during The final integration will start at AdOptica
actuators, which will deform and change ELT operation one set is integrated on once the reference structure has been
the shape of the six petals. M4, while the other is being recoated). delivered. Given the number of compo-
The reference body manufacturing also nents that need to be assembled to a
The 2.7-metre diameter lightweight began in 2017 and six segments have high degree of precision, the integration
structure is made of Boostec® silicon been brazed in the last few months. The will be a lengthy task requiring proce-
carbide, one of the stiffest materials reference surface will need to be lapped dures to ensure that the assembly and
available (stiffer than steel, carbon fibre to 5 microns flatness before being deliv- calibration meet requirements. It should
or beryllium). Its surface has more than ered to Italy. take 1.5 years to fully integrate the M4
5000 holes which will hold the actuators mirror and start the final calibration of
(see Figure 4), while the back surface is To have a mirror fully tested in Chile by each mirror segment and their associated
composed of several ribs to reinforce the early 2024, AdOptica has to ensure capacitive sensors. A test tower is being
structure. Owing to its large dimensions, the procurement and manufacture of all specially developed to verify and test the
the silicon carbide structure is made of the other components, including all the M4. It will be used in Europe to calibrate
six parts brazed together, similar to the voice coil actuators and more than 60% the M4 unit before being transferred
Herschel primary mirror which was man- of the permanent magnets, which are to Chile where it will be used before the
ufactured more than a decade ago. The already in house and are waiting to be mirror is installed on the telescope and
manufacture of the structure is signifi- integrated. In addition, more than half of kept on-site for any future major mainte-
cantly challenging, not only because of the electronics boards are either ready nance activities that may be required.
the depth, length, and thickness of the or under calibration, and most of the
ribs, but also given the requirements on mechanical parts are ready, including the
its straightness, as well as the number reference structure cell support and its
and accuracy of the actuator holes. whiffletree.
Markus Kasper 1 ESO, in collaboration with the Break- ESO 3.6-metre telescope at La Silla,
Robin Arsenault 1 through Initiatives, has modified the VLT was modified and used to carry out the
Ulli Käufl 1 mid-infrared imager VISIR to greatly acceptance tests of the internal chopper.
Gerd Jakob 1 enhance its ability as a planet finder. It This was followed by a performance eval-
Serban Leveratto 1 has conducted a 100-hour observing uation of the Annular Groove Phase Mask
Gerard Zins 1 campaign to search for low-mass plan- (AGPM) coronagraph with a dedicated
Eric Pantin 2 ets around both components of the optical setup incorporating a line-tunable
Philippe Duhoux 1 binary a Centauri, part of the closest CO2 laser, elliptical mirrors and germa-
Miguel Riquelme 1 stellar system to the Earth. Using adap- nium lenses. Four AGPM coronagraphs
Jean-Paul Kirchbauer 1 tive optics and high-performance coro- were tested, three specifically optimised
Johann Kolb 1 nagraphy, the instrument reached for the NEAR filter (10–12.5 μm) and an
Prashant Pathak 1 unprecedented contrast and sensitivity older sample manufactured in 2012 and
Ralf Siebenmorgen 1 allowing it to see Neptune-sized planets optimised for wavelengths between 11
Christian Soenke 1 in the habitable zone, if present. The and 13.1 μm. Surprisingly, the older coro-
Eloy Fuenteseca 1 experiment allowed us to characterise nagraph performed best, with a rejection
Michael Sterzik 1 the current limitations of the instrument. ratio of up to 400 at 10.5 μm, and a con-
Nancy Ageorges 3 We conclude that the detection of trast level of < 10 – 4 at 3 λ/D.
Sven Gutruf 3 rocky planets similar to Earth in the
Dirk Kampf 3 habitable zone of the a Centauri System After passing Provisional Acceptance
Arnd Reutlinger 3 is already possible with 8-metre-class Europe (PAE) in November 2018, the
Olivier Absil 4 telescopes in the thermal infrared. NEAR hardware was shipped to Paranal.
Christian Delacroix 4 At the same time, VISIR was dismounted
Anne-Lise Maire 4 from UT3 (Melipal) and brought to Para-
Elsa Huby 5 From an idea to the telescope nal’s New Integration Hall (NIH) in prepa-
Olivier Guyon 6, 7 ration for the on-site installation starting
Pete Klupar 7 The a Centauri system is uniquely suited in early January 2019. As expected, three
Dimitri Mawet 8 to the search for signatures of low- cool-downs of VISIR were required to
Garreth Ruane 8 mass planets in the thermal infrared. The successfully implement all the new modi-
Mikael Karlsson 9 N-band at around 10 μm is best suited fications. First, the aperture wheel was
Kjetil Dohlen 10 for such observations, because this rearranged with the help of the Paranal
Arthur Vigan 10 is where a planet with a temperature like mechanical workshop to include two new
Mamadou N’Diaye 11 Earth’s is brightest. The a Centauri AGPMs and a special optical mask
Sascha Quanz 12 binary consists of the solar-type stars (ZELDA, N’Diaye et al., 2014) to measure
Alexis Carlotti 13 a Centauri A and B, and the planet- and pre-compensate optical aberrations
hosting (Anglada-Escudé et al., 2016) in the instrument. New Lyot filters were
M-dwarf star Proxima Centauri. In a mounted and mechanically centred with
1
ESO previous Messenger article (Kasper et al., the cold stop of VISIR to an accuracy of
2
AIM, CEA, CNRS, Université Paris- 2017), we provided details of how we better than 175 μm (i.e., 1% of the pupil
Saclay, Université Paris Diderot, planned to modify the existing VISIR diameter). The internal chopper, the
Sorbonne Paris Cité, Gif-sur-Yvette, instrument to conduct the necessary wavefront sensor arm and the calibration
France observations with the Very Large Tele- unit were installed with the help of the
3
Kampf Telescope Optics (KT Optics), scope (VLT). This article describes how contractor KT Optics, and all units were
Munich, Germany VISIR was moved to UT4, the innova- successfully tested. In particular, the
4
University of Liège, Liège, Belgium tions and new technologies that were alignment of the calibration unit, which
5
Observatoire de Paris-Meudon, France implemented and how they work, con- uses an elliptical mirror with an aberration-
6
Subaru Telescope, Tokyo, Japan cluding with the execution of the NEAR free field of view of around 0.1 mm in
7
Breakthrough Initiatives, Mountainview, (New Earths in the a Centauri Region) diameter was laborious and required
USA experiment — a unique 100-hour obser- some modifications of the mechanical
8
Caltech, Pasadena, USA vation of the a Centauri system, which mounts on-site.
9
Uppsala University, Sweden took place in early June 2019.
10
L aboratoire d’Astrophysique Marseille, Following the completion of the assembly
France Three years were needed to develop the integration and verification (AIV) activities,
11
Observatoire de la Côte d’Azur, Nice, NEAR experiment from the initial idea, VISIR was transported and mounted to
France from the Phase A review held in July 2016 UT4 (Yepun) in mid-March 2019 (see Fig-
12
Eidgenössische Technische Hochschule to the observing campaign in June 2019. ure 1). After measuring the expected
Zürich, Switzerland Between January and July 2018, ESO’s residual misalignment between the instru-
13
Institude de Planétologie et d’Astro mid-infrared detector test facility Thermal ment and telescope pupil on-sky on
physique de Grenoble, France Infrared MultiMode Instrument (TIMMI2), 24 March, VISIR was taken off the tele-
a decommissioned instrument from the scope again for adjustment by tilting
ESO/NEAR Collaboration
mounted on UT4 and
of the wavefront sensor arm. On-sky
ready for NEAR. The
commissioning started on 3 April 2019 alternative altitude cable
and lasted for 10 half-nights, during wrap connecting the
which the various new functions were instrument to the elec-
tronics racks and helium
tested, and operational procedures were
compressors on the
tuned. a zimuth platform can be
seen on the left hanging
down from the mirror
cell.
Technical innovations, observing modes
and performance
x
y 8
100
0.05 7
Time/night (h)
0.00 5
60
4
– 0.05 3 40
2
– 0.10 20
1
0 0
0.25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Relative field selector position (arcseconds)
Contrast
10 –3 10 –5 a function of radial distance to the
centre.
Affected strongly
by ADI residuals
10 –4 10 –6
0 2 4 6 8 10
Separation (arcseconds)
with a factor of 80, which degraded to out the mini-warmup and cool-down in to the position between the two off-axis
only a factor of 40 about a week into the the morning and be back in business in PSFs and binned the surviving 76 hours
campaign. Somewhat frustratingly, none less than 10 hours, sufficiently quick to of data to 1-minute time resolution, which
of the typical external effects that degrade be ready for the following observing is short enough to avoid any noticeable
coronagraph efficiency (for example, Lyot night. And it was a success! The corona- smearing of the images because of field
stop misalignments or optical aberra- graphic rejection recovered with each rotation. This procedure compressed the
tions) could explain the shape of the mini-warmup, and starting from 1 June full campaign into ~ 4600 frames or 3 Gb.
residual image that we observed. It really we repeated this procedure approximately
looked like an intrinsic degradation of the every three days. A relatively simple high-contrast imaging
coronagraphic mask itself. analysis can help evaluate the detection
limits reached during the campaign. By
Could air, entering the cryostat through The data and preliminary results sorting all the frames according to their
a known tiny leak, freeze out on the parallactic angle, we run a PSF calibration
20-Kelvin cold coronagraphic mask and The campaign data were taken at a procedure based on principal component
produce the loss of contrast? A back-of- VISIR/NEAR detector frame rate of 166 Hz, analysis using all the frames, i.e., pro-
the-envelope estimate showed that such i.e., the detector integration time (DIT) cessing the campaign as a whole rather
a leak could indeed build up an ice layer was 6 ms. Chopping ran at 8.33 Hz for than night-by-night. The calibrated images
of a few microns thickness every day. most of the campaign, and each chop- are then combined using noise-weighted
With the refractive index of solid nitrogen, ping half-cycle thus lasted 60 ms. During averages in order to properly take into
the main constituent of air, ice partly this time span, 48 ms or 8 DITs were account the rather large variations in the
entering the grooves of the AGPM coro- averaged into a single frame, and 12 ms sky background.
nagraphic mask could change the opti- or 2 DITs were skipped for the transition
cal depth of the grooves sufficiently to of the DSM between the two chopping Figure 5 shows the result of this simple
degrade the performance. positions. Each 30-second data file data reduction and the contrast sensitivity
consists of 500 half-cycle frames, and achieved. The 5σ background-limited
So, how were we to test this theory, and the 100 hours of data add up to 6 million sensitivity far away from the star is of the
even more importantly, fix it during the frames or 6 Tb. order of 100 μJy, which is consistent
campaign as it was running? Solid nitro- with our initial goals. At ~ 1.1-arcsecond
gen starts to sublimate at a sufficiently Before entering advanced high-contrast separation, i.e., at the angular size of
high rate to de-ice the coronagraph mask imaging data reduction, some pre-pro- the habitable zone around α Centauri A,
at temperatures that are only moderately cessing was necessary to remove bad the sensitivity is reduced to about 250 μJy
higher than the nominal 20 Kelvin. It frames and reduce the data volume to a mostly by the central glow of the AGPM.
turned out that the temperature after the more manageable size. We removed This does not yet mean that a point
first stage of the instrument warmup, frames with extremely high or variable source can readily be detected at this
lasting just a few hours, is 35–40 Kelvin. background produced, for example, by level, but first estimates using a fake
Tricking the PLC-controlled system into thin clouds or low encircled energy for injected source show that a planet of
stopping the warmup sequence after the the off-axis stars during ineffective AO ~ 350 μJy brightness corresponding to a
first stage and going into cooling again correction, and frames with low corona- temperate Neptune could indeed be seen.
was risky (a glitch could have resulted in graphic suppression through bad cen-
a full warmup which would have taken tring of the PSF on the coronagraph No planet candidate of the size of Neptune
out VISIR for several days), but it paid mask. Finally, we cropped the images to or larger was found in the data so far.
off. A procedure was developed to carry 400 × 400 pixels, carefully centred them While we were obviously hoping for a
Magda Arnaboldi 1 ESO Public Surveys: overview of sis, including the timeline for the delivery
Nausicaa Delmotte 1 engagement rules and status of science data products over the entire
Dimitri Gadotti 1 duration of the survey project. The
Michael Hilker 1 By design, the ESO Public Surveys cover approval of the survey management plan
Gaitee Hussain 1 a variety of research areas, from the is further confirmed by the agreement
Laura Mascetti 2 detection of planets via micro-lensing, signed between ESO’s Director General
Alberto Micol 1 through stellar variability and evolution, the (DG) and the Principal Investigator (PI) of
Monika Petr-Gotzens 1 Milky Way and Local Group galaxies, to each survey.
Marina Rejkuba 1 extragalactic astronomy, galaxy evolution,
Jörg Retzlaff 1 the high-redshift Universe and cosmol- The agreement between the DG and the
Chiara Spiniello 1, 3 ogy. Differently from Large Programmes, PI specifies the milestones for the data
Bruno Leibundgut 1 these projects are planned to span more releases and their content and responsi-
Martino Romaniello 1 than four semesters and last for many bility for the scientific quality and accu-
years. For example, the latest call for the racy of the data products, which is to be
Cycle 2 survey projects for the Visible warranted by the Public Survey team
1
ESO and Infrared Survey Telescope for Astron- under the leadership of the PI. The agree-
2
Terma GmbH, Darmstadt, Germany omy (VISTA) required them to span a time ment states that a final release which
3
Astronomical Observatory of interval of more than three years. These includes the reprocessing of the entire
Capodimonte, Naples, Italy a survey projects all have a legacy value for data set is expected upon completion of
the community at large in addition to the data acquisition for each survey. This final
science goals identified by the proposing data release should take place within
This report on the status of the ESO teams. one year of completion of the data acqui-
Public Surveys includes a brief overview sition for any survey. The PSP was set up
of their legacy value and scientific to periodically review the progress of the
impact. Their legacy is ensured by their ESO science policies for Public Surveys surveys and to assess compliance with
homogeneity, sensitivity, large sky the specification of the survey products.
coverage in multiple filters, large num- The selection of ESO Public Surveys is In May 2019, a PSP review took place to
ber of targets, wavelength coverage a two-step process which starts with the evaluate the scientific impact of the active
and spectral resolution, which make submission of letters of intent. On the Public Surveys.
them useful for the community at large, basis of these letters, the Public Survey
extending beyond the scientific goals Panel (PSP) formulates a coherent, well-
identified by the survey teams. In May balanced scientific programme that takes Operations for ESO Public Surveys
2019, as almost all first-generation into account any synergies among teams
imaging and spectroscopic surveys in the community and the international The ESO Public Survey observations —
completed their observations and second- survey projects. The PSP then provides whether in service mode or visitor mode
generation imaging surveys got well recommendations to ESO including a list — are carried out according to the pro-
underway, the Public Survey Panel of the teams that should be invited to cess defined by the ESO Data Flow Sys-
reviewed the scientific impact of these submit full proposals on the basis of the tem. The raw data acquired for ESO
projects. The review was based on a ranking of the descriptions of their sci- Public Surveys are immediately public.
quantitative assessment of the number ence projects as provided in the letters of Once the Public Survey teams have car-
of refereed publications from the survey intent. In so doing, the PSP fosters active ried out data reduction to remove instru-
teams and archive users. It included collaborations within the community by mental signatures, calibrate the data and
the number of citations, the number of asking independent teams to join, encour- complete the measurements defined by
data releases and statistics on access aging them to optimise science goals their scientific goals, ESO assists the
to archive data by the user community. and observing strategies, and sharing survey teams to define and package their
The ESO Users Committee also dis- resources. data products in a manner consistent
cussed the availability and usage of with the ESO Science Archive and Virtual
ESO Public Survey data by the commu- Once the proposals have been recom- Observatory standards and in agreement
nity during their yearly meeting in April mended for approval by the PSP and with the specifications in the survey
2019. We describe the status of these the Observing Programmes Committee, management plans. The goal is to inte-
projects with respect to their observing data acquisition for each ESO Public grate science data products from the
plans, highlight the most recent data Survey starts. This involves the review Public Surveys into the ESO Archive,
releases and provide links to the result- and assessment of each survey manage- together with the entire archive content
ing science data products. ment plan by the ESO Survey Team. from the La Silla Paranal Observatory.
The survey management plan is an This is done via the Phase 3 process,
essential tool for the survey team, as well which is an audit process that certifies
as for operations at ESO; it details the the integrity, consistency and data quality
data acquisition plan, and the allocated of the products available from the ESO
resources for data processing and analy- Archive and ensures a homogeneous
20 2-1
20 3-1
20 4-1
20 5-1
20 17-1
20 6-1
20 8-1
20 - 0
20 1-0
20 2-0
20 - 0
20 3 - 0
20 4 - 0
20 4 - 0
20 - 0
20 5 - 0
20 - 0
20 - 0
20 - 0
20 6 - 0
20 7-0
20 8 - 0
-0
1
20 2-1
20 9-1
20 1-1
20 3-1
20 4-1
20 5-1
20 6-1
20 8-1
20 - 0
20 - 0
20 - 0
20 - 0
20 - 0
20 - 0
20 - 0
20 - 0
20 - 0
20 - 0
-0
VST Survey ID Science Area (square Filters Magnitude limits Total time
degrees) (hours)
KiDS — Kilo-Degree Survey Extragalactic 1350 b u’ g’ r’ i’ 24.1 24.6 24.4 3421
http://kids.strw.leidenuniv.nl/ 23.4
(de Jong et al., 2013)
ATLAS Wide area/baryon 4700 c u’ g’ r’ i’ z 22.0 22.2 22.2 1585
http://astro.dur.ac.uk/Cosmology/vstatlas/ acoustic oscillations 21.3 20.5
(Shanks et al., 2013)
VPHAS+ — VST Photometric Hα Survey of the Southern Stellar astrophysics 1800 d u’ g’ Hα r’ i’ 21.8 22.5 21.6 1200
Galactic Plane 22.5 21.8
http://www.vphas.eu (Drew et al., 2013)
VISTA Cycle 1 Science Area (square Filters Magnitude limits Total time
degrees) (hours)
UltraVISTA Deep high-z 1.7 Deep Y J H Ks 25.7 25.5 25.1 1832
http://home.strw.leidenuniv.nl/~ultravista/ 0.73 Ultra deep NB118 24.5 26.7 26.6
(McCracken et al., 2013) 26.1 25.6 26.0
VHS — VISTA Hemisphere Survey Southern sky 17 800 Y J H Ks 21.2 21.1 20.6 4623
http://www.ast.cam.ac.uk/~rgm/vhs/ 20.0
(McMahon et al., 2013)
VIDEO — VISTA Deep Extragalactic Observations Survey Deep high-z 12 Z Y J H Ks 25.7 24.6 24.5 2073
https://www.eso.org/sci/observing/PublicSurveys/ 24.0 23.5
sciencePublicSurveys.html
(Jarvis et al., 2013)
VVV — VISTA Variables in the Via Lactea Milky Way 560 Z Y J H Ks 21.9 21.1 20.2 2205
http://vvvsurvey.org/ 18.2 18.1
(Hempel et al., 2014)
VIKING — VISTA Kilo-Degree Infrared Galaxy Survey Extragalactic 1500 Z Y J H Ks 23.1 22.3 22.1 2424
http://www.astro-wise.org/ 21.5 21.2
(Edge et al., 2013)
VMC — VISTA Magellanic Clouds Survey Resolved star 180 Y J Ks 21.9 21.4 20.3 2047
http://star.herts.ac.uk/~mcioni/vmc/ formation history
(Cioni et al., 2013)
VISTA Cycle 2 Science Area (square Filters Magnitude limits Total time
degrees) (hours)
VINROUGE* — Kilonova counterparts to gravitational wave Kilonova 300 Y J Ks 21.0 21.0 20.1 77
sources counterparts to
http://www.star.le.ac.uk/nrt3/VINROUGE/ GW sources
(Tanvir et al., 2017)
Cont. UltraVISTA — Completing the legacy of UltraVISTA High-z 0.75 J H Ks 26.0 25.7 25.3 567
http://home.strw.leidenuniv.nl/~ultravista/
VVVX* — Extending VVV to higher Galactic latitudes Milky Way 1700 J H Ks Ks = 17.5 1631
http://vvvsurvey.org/
VEILS* — VISTA Extragalactic Infrared Legacy Survey Galaxy evolution, 9 J Ks J < 23.5 847
http://www.ast.cam.ac.uk/~mbanerji/VEILS/veils_index.html AGN, SN Ks < 22.5
G-CAV — Galaxy Clusters At VIRCAM Galaxy clusters 30 Y J Ks 24.5 24 23 440
http://www.oats.inaf.it/index.php/en/2014-09-12-12-59-22/
tematiche-di-ricerca/macroarea-1-en/670-galaxy_cluster.html
VISIONS* — VISTA star formation atlas Star formation 550 J H Ks 21.5 20.5 19.5 449
https://visions.univie.ac.at atlas
SHARKS — Southern Herschel-Atlas Regions Near-infrared 300 Ks 22.7 929
Ks-band survey counterparts for
https://www.iac.es/sharks/ radio sources
Table 1. (Upper) VLT Survey Telescope Public Sur- Table 2. (Centre) Cycle 1 VISTA Public Surveys; Table 3. (Lower) Cycle 2 VISTA Public Surveys
veys. These projects began operations in October these projects began operations in April 2010 and began operations in April 2017. The four Cycle 2
2011 and data acquisition is now completed accord- are now all completed but for the VHS subareas VISTA surveys that explore the time domain are
ing to their survey management plans. The total close to the South Galactic Pole. The total number indicated by an asterisk in the table. The total
number of completed hours is reported to the 30 of completed hours is reported to the 30 September number of completed hours by 30 September 2019
September 2019 date. 2019 date. is shown in the last column.
O
G
e
+
S
C
AS
A
EO
TO
LS
AC
ES
ug
The ESO library routinely monitors
ST
AS
V/
D
VH
VM
N
DE
SS
L
KI
G
D
KI
VV
Ro
a–
VI
AT
H
VI
LE
VI
N
PE
tr a
VP
ai
N
VA
T
VI
Ul
VS
scientific networks (for example, of the Small Magellanic Cloud. All data of files downloaded by the community for
PESSTO–WISeREP). releases were promptly advertised via the each ESO Public Survey. The lower chart
Archive/Phase 3 web pages, followed up shows the numbers of catalogues, the
All survey teams have successfully with specific announcements on the ESO numbers of distinct users and the num-
published several data releases for some science page, the Science Newsletter 9 bers of queries carried out using the ESO
of their science data through the Phase 3 and the ESO archive community forum10. catalogue query interface 12 to access
process (Arnaboldi et al., 2014); an over- ESO Public Survey catalogues. On aver-
view of these releases is available via The most recent data releases join a age, users of the ESO catalogue query
this webpage8. Since January 2019, the large number of data collections from the interface carry out at least 21 independ-
total volume of science data products ESO Public Surveys that can be browsed ent queries to access catalogue records.
released from the ESO Public Surveys using the Archive Science Portal11. The
amounts to 27.4 Tb, including ancillary science data products from the ESO An enhanced archive capability allowing
files. The data releases published this Public Surveys amount to a total volume programmatic access13 results in anony-
year include: the fourth data release of of 68.6 Tb (nearly 8.5 × 105 files) which mous exploration and retrieval of cata-
KIDS (> 1000 square degrees) and are currently accessible via the ESO logue records (and other products) via
UltraVISTA (deep stacked images of the Archive. The science data products that Virtual Observatory tools, for example,
COSMOS field from observations can be actively queried and downloaded Aladin and Topcat. This new service
acquired between December 2009 and amount to nearly 320 000 catalogue files, allows users to repeat queries in an auto-
June 2016); the proper motion of selected half a million astrometrically and photo- mated fashion, in order to perform more
stars in the Milky Way disc and bulge metrically calibrated images, and 56 000 complex queries by combining data from
from the VVV near-infrared Astrometric 1D extracted spectra. In Figure 4 we different surveys or other content of the
Catalogue (VIRAC); accurate PSF-fitting show a collection of on-sky footprints of ESO Science Archive, thereby enhancing
photometry of the 300 square degrees the data releases published during the the scientific use of the catalogue content
around the Galactic centre; and the fifth last year by the ESO Public Survey teams. of the ESO Archive. One interesting
data release of VMC with full coverage statistic is the number of distinct users —
1583 users from 77 different countries —
Data download statistics who have downloaded ESO Public
Figure 4. Montage of the footprints of the data
releases from the ESO Public Surveys published by
Survey science products published via
2019, as shown on the ESO Archive Science Portal In Figure 5 we show the cumulative curves the ESO Archive. To place this in context,
interactive interface. of the data volume (Gb) and the number the fraction of distinct users who access
16 000
12 000
Data volume (Gb)
Number of files
150 000
10 000
8000
100 000
6000
4000
50 000
2000
0 0
12–12
12–11
03–12
06–12
09–12
03–13
06–13
09–13
12–13
03–14
06–14
09–14
12–14
03–17
06–17
09–17
12–17
03–18
03–15
06–18
09–18
12–18
06–15
09–15
12–15
03–16
06–16
09–16
12–16
03–19
09–19
12–17
06–18
12–18
03–19
09–19
12–11
06–12
12–12
06–13
12–13
06–14
12–14
06–15
12–15
06–16
12–16
06–17
VHS VVV/3 UltraVISTA VMC VIDEO VHS VVV/3 UltraVISTA VMC VIDEO
VIKING VPHAS+ KiDS V-ATLAS Gaia-ESO VIKING VPHAS+ KiDS V-ATLAS Gaia-ESO
PESSTO Lega-C Vandels G-CAV VISIONS PESSTO Lega-C Vandels G-CAV VISIONS
data products from ESO Public Surveys Figure 5. Charts show- 2500
ing the cumulative data
is 46% of the total number of users
volume (upper left) and
(3457) who have downloaded science
Number of catalogues/users/queries
O 3
Ju 4
Ju 5
Ju 6
Ju 7
Se 8
19
M 3
O 4
4
15
16
17
–1
–1
–1
–1
–1
–1
–1
n–
p–
n–
n–
p–
n–
ar
ar
ct
ct
ec
ec
ec
ec
ec
Se
D
D
D
D
the most reliable/useful documentation to announced 2019 a call for letters of intent Drew, J. E. et al. 2013, The Messenger, 154, 41
Edge, A. et al. 2013, The Messenger, 154, 32
support the scientific use of the available for 4MOST Community Surveys on
Hempel, M. et al. 2014, The Messenger, 155, 29
data products. 26 November 2019. The community sur- Jarvis, M. J. et al. 2013, The Messenger, 154, 26
veys will access up to 30% of the availa- Liske, J. & Mainieri, V. 2019, The Messenger, 177, 61
ble time on the 4MOST spectrograph McCracken, H. J. et al. 2013, The Messenger, 154, 29
McLure, R. et al. 2017, The Messenger, 167, 31
The Public Survey Panel review on VISTA over a period of five years.
McMahon, R. et al. 2013, The Messenger, 154, 35
These projects will complement the GTO Randich, S. et al. 2013, The Messenger, 154, 47
Membership and evolution of the Public surveys that were presented during the Shanks, T. et al. 2013, The Messenger, 154, 38
Survey Panel 4MOST workshop at ESO in May and Smartt, S. et al. 2013, The Messenger, 154, 50
Sutherland, W. et al. 2015, A&A, 575, 27
are described in the special issue of the
Tanvir, N. et al. 2017, ApJ, 848, 27
In 2005, the Public Survey Panel (PSP) Messenger in March 2019 7. van der Wel, A. et al. 2016, The Messenger, 164, 36
was set up as a subpanel to the Observ-
ing Programmes Committee, its role Large surveys are considered a key tool
Links
being later extended to include monitor- in observational astronomy because
ing of the progress of the Public Survey they allow explorations that go beyond 1
ESO Public Surveys Project webpage:
projects. Four chairs organised the work individual targeted observations. They are http://www.eso.org/sci/observing/PublicSurveys/
of the panel: Duccio Macchetto, Simon characterised by large investments that sciencePublicSurveys.html
2
ESO Telescope Bibliography telbib:
White, Danny Lennon, and Miguel Mas comprise dedicated telescopes and
http://telbib.eso.org
Hesse (the current chair). From 2005 to instruments, a wide community of astron- 3
The Cambridge Astronomy Survey Unit:
2012, ESO set up three separate panels: omers involved in the science projects http://casu.ast.cam.ac.uk/
the VISTA, VST imaging, and Spectro- and extended networks for the data distri- 4
The Wide Field Astronomy Unit:
http://www.roe.ac.uk/ifa/wfau/
scopic Public Survey Panels; the VISTA bution and analysis. The scientific success 5
TERAPIX: http://terapix.iap.fr/
and VST Panels were later merged. Since of such survey projects includes the leg- 6
CALET: https://calet.org/
2018 there has been only one panel for acy value of science products that are 7
Astro-WISE: http://www.astro-wise.org/
all ESO Public Surveys, whether imaging made available through the archives for 8
ESO Phase 3 Data Releases: http://eso.org/rm/
publicAccess#/dataReleases
or spectroscopy. In the near future, the further scientific exploration by the com- 9
ESO Science Newsletter: http://www.eso.org/sci/
PSP will advise ESO on the selection of munity. ESO Public Surveys are an exam- publications/newsletter/
4MOST surveys (see de Jong et al., 2019). ple of an effective implementation of this 10
ESO Archive Community Forum:
strategy, with the goal of supporting the https://esocommunity.userecho.com/
11
ESO Archive Science Portal:
scientific advancement of its community.
https://archive.eso.org/scienceportal/
The PSP review in May 2019 12
ESO catalogue query interface:
https://www.eso.org/qi/
As most of the Public Surveys completed Acknowledgements 13
ESO Archive Programmatic Access webpage:
http://archive.eso.org/programmatic/
their data acquisition in 2019, the goal The authors would like to thank colleagues at the 14
VISTA image of LMC:
and objectives of the PSP review held in La Silla Paranal Observatory for their work, and sup- https://www.eso.org/public/news/eso1914/
May at the ESO headquarters were to port by the science operations of the ESO Public 15
VISTA image of Milky Way bulge:
assess the scientific success of the Public Surveys. They wish to acknowledge the colleagues https://www.eso.org/public/news/eso1339/
from the ESO Directorate of Engineering who sup- 16
VISTA image from KIDS survey:
Surveys using criteria that included pub- ported the development of the tools required for car- https://www.eso.org/public/news/eso1642/
lished results in refereed journals, the rying out Phase 1, Phase 2 and Phase 3 operations 17
First Light from Gravitational Wave Source:
progress of the analysis, potential scien- for ESO Public Surveys and the Archive science https://www.eso.org/public/news/eso1733/
tific extensions, complementarity with interfaces. The authors thank the ESO library team 18
Messenger issue dedicated to 4MOST GTO
for their careful monitoring of the refereed publica- surveys: https://www.eso.org/sci/publications/
other telescopes, activities to promote tions. The authors wish to thank the PIs of the Public messenger/archive/no.175-mar19/messenger-
the surveys, and the use of survey data Surveys and their collaborators, including the data no175.pdf
products for independent projects in the centres at CASU3, WFAU4, Astro-WISE14 (Astro
community. The PIs of the twenty active nomical Wide-field Imaging System for Europe) and
TERAPIX5, for their hard work and support of the Notes
Public Surveys were invited to the review ESO mission. Finally, ESO wish to thank the PSP
and asked to address the above criteria chairs and the colleagues in the community who a
strofit2 Marie Skłodowska-Curie Fellow
A
during their presentations. All PIs but one served as members in the different PSPs for their b
T he survey area was reduced from the original
attended the review and contributed to efforts and their support with the definition of how approved coverage to coincide with the VIKING
to execute the survey programmes. footprint.
a lively and constructive discussion. The c
T he survey area was increased to 4700 square
PSP report was presented to the Scien- degrees after approval by the PSP; data-taking is
tific Technical Committee at its meeting in References currently active.
October 2019. d
T he survey area was reduced from that approved
Arnaboldi, M. et al. 2007, The Messenger, 127, 28 by the PSP following a request by the PI to end the
Arnaboldi, M. et al. 2014, The Messenger, 156, 24 survey early.
Arnaboldi, M. et al. 2017, The Messenger, 168, 15
Outlook and Conclusions Capaccioli, M. & Schipani, P. 2011, The Messenger,
146, 2
Cioni, M.-R. et al. 2013, The Messenger, 154, 23
While work on the scientific analysis con- de Jong, R. 2019, The Messenger, 175, 3
tinues for the twenty Public Surveys, ESO de Jong, R. et al. 2013, The Messenger, 154, 44
log g
Lorenzo Morelli 4 Le Borgne et al., 2004) and the X-shooter
Spectral Library (XSL; Chen et al., 2014).
The latter showcases the problems that
1
ESO increasing resolution and multi-order 4
2
Dipartimento di Fisica e Astronomia cross-dispersed spectrographs bring;
“G. Galilei”, Università di Padova, Italy synthetic broadband optical (UBV) colours
3
INAF–Osservatorio Astronomico di show poor agreement with observed
Padova, Italy colours from the Bright Star Catalogue 4.5 4 3.5
4
Instituto de Astronomía y Ciencias (on average at ~ 7%; see Table 5 and Fig- log Teff
Planetarias, Universidad de Atacama, ure 26 in Chen et al., 2014). The differ-
Copiapó, Chile ences are likely related to pulsating varia-
ble stars observed at different phases. 6
Slit losses are another issue; for many 4
N
Empirical stellar spectral libraries have stars these are caused by the attenuation
2
applications in both extragalactic and of flux, or other losses inherent to slit-
stellar studies. We have assembled the based spectrographs. 0
MUSE Spectral Library (MSL), consist- O B A F G K M C/S
ing of 35 high-quality spectra of stars We embarked on a project to build an Spectral type
covering the Hertzsprung–Russell dia- empirical spectral library without slit
gram, and verified the continuum shape losses using the MUSE (Multi-Unit Spec- Figure 1. Properties of the MSL stars. Top: surface
gravity log g vs. effective temperature Teff for stars with
of our spectra with synthetic broadband troscopic Explorer; Bacon et al., 2010)
[Fe/H] ≤ −0.5 dex (crosses), −0.5 < [Fe/H] < 0.0 dex
colours. We also report indices from IFU, with the goal of spanning all of the (open circles), and [Fe/H] ≥ 0.0 dex (filled circles).
the Lick system, derived from the new major sequences on the Hertzsprung– Bottom: distribution of the stars by spectral type.
observations. Our data demonstrate Russell diagram and serving as a bench-
that integral field units (IFUs) are excel- mark for the shapes of other theoretical
lent tools for building spectral libraries and empirical spectra. Our final products Data reduction was performed with the
with reliable continuum shapes that can are spectra suitable for galactic model- ESO MUSE pipeline (v. 2.6) within the
be used as templates for extragalactic ling, stellar classification and other appli- ESOReflex 3 environment (Freudling et
studies. cations. Here we report on our first sam- al., 2013). The 1D spectra were extracted
ple of 35 MSL spectra. using a circular aperture with a radius of
6 arcseconds. This number was selected
Introduction and sample Our initial sample numbered 33 XSL stars1. after experiments with different aperture
In addition, HD 193256 and HD 193281B sizes, to guarantee that “aperture” losses
Empirical stellar spectral libraries are a serendipitously fell inside the MUSE field led to less than a 1% change in the over-
universal tool in modern astronomy, with of view. The full sample is described in all slope of the spectra from the blue to
applications in both extragalactic and Table 1 of Ivanov et al. (2019). the red.
galactic stellar studies. They can have
multiple uses: to match and remove Three stars were treated differently,
continua to reveal weak emission lines; Observations and data reduction without major loss of continuum fidelity.
as templates to measure stellar kinemat- For the asymptotic giant branch star
ics in galaxies; and to measure stellar The spectra were obtained with MUSE [B86] 133 we reduced the extraction
parameters such as effective tempera- at the European Southern Observatory aperture radius to 4 arcseconds to avoid
tures and surface gravities. Theoretical (ESO) Very Large Telescope, Unit Tele- contamination from nearby sources. For
stellar models can have significant weak- scope 4 (Yepun), on Cerro Paranal, Chile. HD 193256 the aperture had a radius
nesses; for example, Sansom et al. (2013) Table A.1 in Ivanov et al. (2019) gives the of 4.6 arcseconds, and the sky annulus
found discrepancies in Balmer lines and observing log. We obtained six exposures had an inner radius of 4.6 arcseconds
the incomplete treatment of molecules for each target, except for HD 204155 and a width of 2 arcseconds because the
(also shown by Castelli, Gratton & which was observed 12 times. We placed star was close to the edge of the MUSE
Kurucz, 1997). This occasionally leads to the science targets at the same spaxels field of view. HD 193281 is a binary with a
poorly predicted broad-band colours. At as the spectrophotometric standards to separation of ~ 3.8 arcseconds. We dis-
the same time, the typical empirical minimise any residual systematics from entangled the two spectra as described
libraries suffer from low resolution and/or the instrument. in Ivanov et al. (2019).
12
10
HD 057060 07e
10 HD 096446 B2III
8
HD 174966 A3
HD 167278 F2
HD 163810 G3V 6
HD 167278 K3V
5
4
(B86) 133 M4
0 HD 067507 CNv 0
6000 8000 10 4 4850 4900 4950 6600 6700 8500 8600 8700
Wavelength (Å) Wavelength (Å) Wavelength (Å) Wavelength (Å)
Figure 2. Comparison of a subset of our MSL spectra fitted second-order polynomials to the University through grants DOR1715817/17,
(black) with the XSL spectra (red; boxcar smoothed DOR1885254/18, DOR1935272/19, and
ratios and extrapolated them over the full
over 8 pixels). The spectra are normalised to unity BIRD164402/16.
between the two vertical dotted lines shown on the wavelength range covered by the XSL
left, and shifted vertically for display purposes. Left: library to demonstrate that, if these
entire MUSE spectral range; right: zoom around the trends hold, the overall peak-to-peak flux References
Hβ, Hα, and Ca triplet features (left to right). No
differences can easily reach ~ 20%,
radial velocity correction is applied. Bacon, R. et al. 2010, Proc. SPIE, 7735, 773508
meaning that the overall continua of the Castelli, F., Gratton, R. G. & Kurucz, R. L. 1997,
cross-dispersed spectra is somewhat A&A, 318, 841
The final MSL spectra have signal-to- ill-defined. Chen, Y.-P. et al. 2014, A&A, 565, A117
Freudling, W. et al. 2013, A&A, 559, A96
noise ratios S/N > 70–200 and are availa-
Ivanov, V. D. et al. 2019, A&A, 629, 100
ble via the ESO MUSE webpage 2 or via Finally, we calculated synthetic Sloan Le Borgne, J.-F. et al. 2003, A&A, 402, 433
CDS/VizieR 3. The Lick indices (Worthey Digital Sky Survey (SDSS) colours from Le Borgne, D. et al. 2004, A&A, 425, 881
et al., 1994) that fall within the wavelength both MSL and XSL spectra (Figure 5 in Pickles, A. J. 1998, PASP, 110, 863
Prugniel, P. & Soubiran, C. 2001, A&A, 369, 1048
range covered by MUSE were measured Ivanov et al., 2019). The MUSE sequences
Sansom, A. E. et al. 2013, MNRAS, 435, 952
in the new MSL spectra (Table C.1 in are slightly tighter than the XSL ones, con- Soubiran, C. et al. 1998, A&AS, 133, 221
Ivanov et al., 2019). firming that the IFU MUSE spectra have Yan, R. et al. 2019, ApJ, 883, 175
more reliable shapes. This is expected in Worthey, G. et al. 1994, ApJS, 94, 687
light of the slit losses and the imperfect
Analysis and discussion order stitching of the XSL spectra. Fur- Links
thermore, X-shooter has three arms and
We demonstrate excellent agreement is in effect three different instruments; 1
he XSL library: http://xsl.u-strasbg.fr/
T
2
T he MUSE spectral library at the ESO MUSE web-
between the 6 (or 12 in the case of some of the colours can mix fluxes from
page: https://www.eso.org/sci/facilities/paranal/
HD 204155) individual observations (Fig- the different arms, which may contribute sciops/tools/MUSE_Spectral_Library.html
ures 2 and A.1 in Ivanov et al., 2019). A to the larger scatter. 3
T he MUSE spectral library at VizieR/CDS: http://
direct comparison of the MSL and XSL cdsarc.u-strasbg.fr/viz-bin/cat/J/A+A/629/A100
spectra for eight randomly selected stars
across the spectral type sequence is Acknowledgements
shown in select wavelength ranges in This paper is based on observations made with
Figure 2. In most cases, the agreement the ESO VLT at the La Silla Paranal Observatory
on a scale of a few hundred pixels — in (Programme ID 099.D-0623). We have made exten-
other words, within the same X-shooter sive use of SIMBAD at the Centre de Données
astronomiques Strasbourg (CDS) and of the VizieR
spectral order — is excellent. However, tool and CDS, Strasbourg, France. Enrico M. Corsini,
on a larger scale we find deviations Elena Dalla Bontà, Lorenzo Morelli, and Alessandro
between the XSL and MSL spectra. We Pizzella acknowledge financial support from Padua
GRAVITY Science
Δ Declination (µas)
10
Differential phase (°)
Normalised flux
0
1.0 2.172
0.2
0.8 –10
0.0 2.170
0.6
Phase – 20 2.168
– 0.2
0.4 BLR model
BLR model
– 0.4 0.2 – 30
2.14 2.15 2.16 2.17 2.18 2.19 2.20 30 20 10 0 –10 – 20 – 30
Observed wavelength (µm) Δ Right ascension (µas)
c Yv d 150 3
PAJet = 222°
Y
100 2
ev
e¦
50 1
H
"KNTC 0 0
e¦
ev
, %+ – 50 –1
MFTK@Q
BKNTC
CDMRHSX
–100 –2
1@CH@K `
BKNTC
CDMRHSX
–150 –3
150 100 50 0 – 50 –100 –150
1 LHM 1 !+1 Q Δ Right ascension (µas)
Figure 1. GRAVITY spatially resolves the broad emis- This is the result of net ordered rotation of the engines. The key components of AGN
sion line kinematics of 3C 273. (a) Paa line profile line-emitting gas. By comparing a kinematic model
are small on the sky, at micro- to milli-
(black) and averaged differential phase (blue), show- of the emission region (c) to GRAVITY data, we
ing non-zero phases and a change of sign across find that a thick disc configuration viewed at low arcsecond scales, requiring long baselines
the broad emission line. (b) Photocentre positions inclination best explains the data (d). The model also at the VLTI and Keck Interferometer. AGN
measured at each line channel, showing a clear sep- provides estimates of the mean emission radius are also relatively faint sources, so far only
aration between red and blue which corresponds and central black hole mass. Adapted from GRAVITY
detected in optical interferometry with
to a velocity gradient at a position angle perpendicu- Collaboration (2018).
lar to the large-scale radio jet of 3C 273 (black line). 8–10-metre-class telescopes and instru-
mentation with excellent sensitivity. Con-
tinuum measurements with the Keck
luminosity, roughly as R ~ L1/2. That rela- campaigns to measure R via an estimate Interferometer (for example, Kishimoto et
tionship can be understood as atomic based on L). Secondary methods so far al., 2011) and the Astronomical Multi-
gas emission being produced under opti- provide all available active galactic nucleus BEam combineR (AMBER) on the VLTI
mal photoionisation conditions (constant (AGN) black hole mass measurements (Weigelt et al., 2012) provide information
received flux). This radius-luminosity rela- in large samples and out to high redshift. about hot dust surrounding the nucleus.
tion allows “secondary” methods for The broad line region (BLR) is even smaller
estimating black hole masses using a Interferometry provides an independent (angular size < 0.1 milliarcseconds [mas])
single optical spectrum (replacing long method for spatially resolving AGN central and is impossible to resolve in standard
European Union’s Horizon 2020 research and inno- FIS/00099/2013, SFRH/BSAB/142940/2018 [P. G.] GRAVITY Collaboration 2018, Nature, 563, 657
vation programme (Grant Agreement No. 743029), and PD/BD/113481/2015; M. F. in the framework of GRAVITY Collaboration 2019a, submitted to A&A,
from the Irish Research Council (IRC Grant: the Doctoral Programme IDPASC Portugal), by NSF arXiv:1910.00593
GOIPG/2016/769) and SFI Grant 13/ERC/12907, grant AST 1909711, by the Heising-Simons Founda- GRAVITY Collaboration 2019b, submitted to A&A
from the Humboldt Foundation Fellowship and the tion 51 Pegasi b postdoctoral fellowship, from the Bentz, M. C. et al. 2013, ApJ, 767, 149
ESO Fellowship programes, from the European Direction Scientifique Générale of Onera and by a Du, P. et al. 2018, ApJ, 856, 6
Research Council under the European Union’s Hori- Grant from Science Foundation Ireland under Grant Grier, C. J. et al. 2017, ApJ, 851, 21
zon 2020 research and innovation programme (Grant number 18/SIRG/5597. Pancoast, A. et al. 2008, MNRAS, 445, 3073
Agreement Nos. 2016–ADG–74302 [EASY], 2015- Kishimoto, M. et al. 2011, A&A, 527, 121
StG-677117 [SFH], 694513, and 742095 [SPIDI]), and Weigelt, G. et al. 2012, A&A Letters, 451, 9
was supported in part by the German Federal Minis- References Zhang, Z.-X. et al. 2019, ApJ, 876, 49
try of Education and Research (BMBF) under the
grants Verbundforschung #05A08PK1, #05A11PK2, Peterson, B. M. et al. 2004, ApJ, 613, 682
#05A14PKA and #05A17PKA, by Fundação para a Kaspi, S. et al. 2000, ApJ, 533, 631
Ciência e a Tecnologia, Portugal (Grants UID/ Rakshit, S. et al. 2015, MNRAS, 447, 2420
DOI: 10.18727/0722-6691/5167
GRAVITY Collaboration (see page 20) Since the first seminal paper addressing Under superb conditions, with seeing
its physical properties (Krolik & Begelman, ~ 0.5 arcseconds and a coherence time
1988), and following numerous observa- of up to 13 ms, it was possible to
The superb resolution of the Very Large tions at many different wavelengths, fringe-track on the nucleus of NGC 1068
Telescope Interferometer (VLTI) and the “torus” concept has evolved and despite its large size and moderate
the unrivalled sensitivity of GRAVITY been modified considerably. At the same brightness. The data obtained were of
have allowed us to reconstruct the first time, increases in computational power excellent quality, with typically < 1% visi-
detailed image of the dust sublimation have facilitated detailed modelling of bility and closure-phase accuracy. The
region in an active galaxy. In the nearby clumpy torus structures. Such models wealth of information provided by the six
archetypal Seyfert 2 galaxy NGC 1068, are consistent with the near- to mid- VLTI baselines has enabled us to recon-
the 2 µm continuum emission traces infrared spectral energy distribution as struct a K-band image based on the
a highly inclined thin ring-like structure well as dust reverberation measurements. obtained closure phases and visibilities
with a radius of 0.24 pc. The observed Observations of almost two dozen with 3-milliarcsecond (mas) resolution.
morphology challenges the picture of a galaxies using the MID-infrared Interfero-
geometrically and optically thick torus. metric instrument (MIDI) on the VLTI have We used the publicly available Multi-
resolved the 1–3 pc scales where warm aperture image Reconstruction Algorithm
dust is responsible for the mid-infrared (MiRA; Thiébaut, 2008) to generate the
Introduction continuum (Burtscher et al., 2013 and ref- image shown in Figure 1, which contains
erences therein). However, measuring a total flux of 155 mJy. The structures
NGC 1068 is one of the best studied the size of the small (< 1 pc) region con- present are robust, having been repro-
nearby active galactic nuclei (AGN), in taining hot dust that emits at near-infra- duced consistently over a wide variety of
which accretion onto a central super- red wavelengths has been possible in parameter settings, and with a signal
massive black hole contributes a signifi- very few galaxies. Also, until GRAVITY level much higher than that expected for
cant fraction of the galaxy’s total luminos- observed NGC 1068, there were no data spurious sources. Full details are in
ity. The observation of broad polarised showing spatial structure in this dust sub- GRAVITY Collaboration (2019).
emission lines by Antonucci & Miller limation region.
(1985) in the nucleus of this Seyfert gal-
axy was central to the development of A new view of NGC 1068
the unified model that explains the differ- Observations and
ences between Seyfert 1 and Seyfert 2 Image Reconstruction The image in Figure 1 is dominated
objects as being due to the presence by knots of continuum arranged in a ring
of a nuclear equatorial structure that both Data on NGC 1068 were obtained in around a central hole, with the south-
obscures and scatters the central emis- November and December 2018 using western side about a factor of two brighter
sion depending on the line of sight. GRAVITY and the four 8-metre UTs. than the north-eastern side. Fitting an
15
Maser disc
(optical) NASA/STScI, (radio) NSF/NRAO/VLA
− 300, 0, 300 km s –1
Po
10 lar
ise
d
ou
t flo
w
co
40.
−5
30.
20.
−10
10.
1 pc
−15 0.
1600 pc 15 10 5 0 −5 −10 −15
Right ascension offset (mas)
ellipse to these knots yields a position the population of masing molecules. In Figure 1. Left: Three-colour image of NGC 1068. The
optical emission is shown in green, the X-ray in red
angle of 50 degrees west of north, an an additional test, we have compared the
and the radio jet in blue. Right: Reconstructed image
inclination of 70 degrees and a radius of spectral energy distributions predicted of the 2 µm continuum (blue colour scale) in the cen-
about 0.24 pc. The size matches remark- by models with the photometry from MIDI tral 2.1 pc of NGC 1068, showing the reconstructed
ably well the expected dust sublimation and GRAVITY. For reasonable parameter beam size in the lower left. The white dashed ellipse,
fitted to the brightest knots, traces a ring that
radius for large graphite grains in the ranges, the models tend to over-predict
matches the expected range of dust sublimation
radiation field of an AGN with an intrinsic the mid-infrared continuum and have radii (orange dotted ellipses). The filled black circle in
bolometric luminosity of ~ 4 × 1045 erg a near-infrared slope that is too shallow. the centre of the ring, denoting the location of the
s –1 as expected for NGC 1068. And if one AGN, has been matched to the kinematic centre
derived from the maser kinematics, and hence fixes
aligns the central hole in the near-infrared As an alternative, we considered whether
the relative position of the maser distribution. The
continuum to the location of the central the mid- and near-infrared continua have radio continuum has been positioned using the
black hole inferred from the maser kine- a common origin at all. Cool (~ 700 K) masers as a coordinate reference. The green dashed
matics (Gallimore &Impellizzeri, 2019), dust behind a screen of extinction pro- lines outline the bipolar ionised outflow also seen in
polarisation data. The grey dashed ellipse indicates
then the positions of the lower-velocity vides an unexpectedly good fit to the
the size of the 10-metre continuum in the MIDI data.
maser spots match up remarkably well spectral energy distribution, including the
with the south-western side of the ring. silicate dip. But the modest AK ~ 0.9
This suggests that the masers and the magnitude extinction is far less than the requires that most of the mid-infrared
hot dust trace a common disc, and hence lower limit of AK ~ 6 magnitudes required continuum originates in a different struc-
that the brighter south-western side of by the non-detection of broad Bra ture on larger scales. Disc-plus-wind
the ring is the near side. This geometry is at 4 µm (Lutz et al., 2000) and the high- models such as those described by
consistent with that implied by the jet and column density implied by the HCN1–0 Hönig (2019) would imply that the other
the ionisation cone, which are oriented emission at 3 mm (García-Burillo et al., structure is in fact the outflow driven by
toward us on the northern side. 2016; Imanishi et al., 2018). the AGN.
°
to that of the maser disc, arguing for
70
i=
a common origin. The structure and pho-
tometry are consistent with dust at
t
igh
~ 1500 K behind AK ~ 5.5 magnitudes of
fs
foreground extinction. This matches what
eo
Lin
is expected from the upper limit to the
broad Bra line, and could originate in the
dense and turbulent gas distribution
observed on scales of 1–10 pc. In such a
scenario, much of the mid-infrared con- s
ud
tinuum would originate in a separate r clo
ase
structure, likely associated with the AGN- M Thic k molecular gas disc
driven outflow.
5
Acknowledgements isc
i nd
t th
See page 23. Ho
1
u st )
MID -IR d 5 (p
c
References R
0. 2
Antonucci, R. R. J. & Miller, J. S. 1985, ApJ, 297, 621
Burtscher, L. et al. 2013, A&A, 558, 149 0
García-Burillo, S. et al. 2016, ApJ, 823, L12
Gallimore, J. & Impellizzeri, V. 2019, submitted to
ApJ Figure 2. Sketch of the observed central structures. from the disc periphery. ALMA observations of HCN
GRAVITY Collaboration 2019, accepted by A&A The K-band emission traces the inner rim of a thin and HCO+ show a turbulent structure, which rotates
Hönig, S. F. 2019, accepted by ApJ disc of hot gas and dust, at or close to the dust sub- in the opposite direction to the maser disc (Imanishi
Imanishi, M. et al. 2018, ApJ, 853, L25 limation radius of 0.24 pc. The inner water masers et al., 2018). The turbulence found in the molecular
Krolik, J. H. & Begelman, M. C. 1988, ApJ, 329, 702 are cospatial with the hot K-band dust. The masers gas structure argues for a thick disc, which contains
Lutz, D. et al. 2000, ApJ, 530, 733 stretch out to 1 pc (Gallimore et al., 2001). Mid-infra- enough gas mass to reach column densities
Thiébaut, E. 2008, Proc. SPIE, 7013, 70131I red observations show warm dust on roughly the that screen the central region from the observer by
same scales as the outer masers, likely originating AK ~ 5.5 magnitudes.
DOI: 10.18727/0722-6691/5168
GRAVITY Collaboration (see page 20) vitational redshift, the most precise mass- It is embedded in hot gas and sur-
distance measurement, the test of the rounded by a cluster of high velocity
equivalence principle, and the detection stars. They buzz around the black
On a clear night, our home galaxy, the of orbital motion near the black hole. hole on trajectories which are, like the
Milky Way, is visible as a starry ribbon behaviour of the hot gas, governed by
across the sky. Its core is located in the gravitational field of the black hole.
the constellation of Sagittarius, approx- The heart of the Milky Way
imately where the bright glow is inter- With GRAVITY we are unravelling what is
rupted by the darkest dust filaments. At the heart of the Milky Way, 26 000 light- happening in the centre of our Galaxy
There, hidden, lies a massive black hole. years from Earth, is Sagittarius A* (Sgr A*, with unprecedented angular resolution.
To peer through the obscuring clouds pronounced “Sag-A-star”), the closest The instrument operates at infrared
and see the stars and gas near the massive black hole to us and, with a wavelengths around 2 microns. GRAVITY
black hole we use GRAVITY. The main lensed angular diameter of 53 microarc combines the light beams of the four
GRAVITY results are the detection of gra- seconds (µas), the largest one on the sky. individual 8.2-metre Unit Telescopes at
4000
0.175 3000
2000
0.125 0
– 1000
Δ Declination (ೀ)
0.1
– 2000
2000 2005 2010 2015 2020
0.075 Time (year)
250
0.05
200
Residuals vLSR (km s –1)
0.025
150
0
100
hole and are consistent with a small What’s next? properties, for example, the sense of
region of heated electrons (a “hot spot”), rotation of the hot gas.
moving in an orbit around the black hole. Continuing observations of S2 are
The GRAVITY observations also revealed expected to reveal a second relativistic
changes in the polarisation angle over effect on the star’s orbit, namely the References
the course of the flare. In particular, as Schwarzschild precession. General rela- GRAVITY Collaboration 2017, A&A, 602, 23
the centroid of the emission region com- tivity predicts that the orbit of S2 is not GRAVITY Collaboration 2018a, A&A, 615, 15G
pletes one orbit around the black hole, a closed Keplerian ellipse but an open GRAVITY Collaboration 2018b, A&A, 618, 15
the polarisation angle also makes a single rosette-like trajectory, where the peri- GRAVITY Collaboration 2019a, Phys. Rev. Lett., 122,
101102
loop. These polarisation measurements apse, i.e., the closest point to the black GRAVITY Collaboration 2019b, A&A, 625, 10
indicate the presence of a strong mag- hole, shifts by a small angle per revolution
netic field in the immediate vicinity of the which rotates the ellipse over time.
black hole and might indicate a magnetic Moreover, studying multiple flares as an
origin of the flare. ensemble will shed light on accretion
DOI: 10.18727/0722-6691/5169
GRAVITY Collaboration (see page 20) rable to the binary orbit, had remained GRAVITY observations of two such
unresolved for a long time because the objects: the hypercritical accretor and
required sub-milliarcsecond spatial reso- exotic microquasar SS433, and the
The GRAVITY instrument at the Very lution is significantly beyond the diffraction wind-accreting high-mass XRB BP Cru.
Large Telescope Interferometer has led limit of even extremely large telescopes.
to the first spatially resolved observa- Resolving these structures is, in fact,
tions of X-ray binaries at scales compa- challenging even for optical interferometry, Resolving super-Eddington outflows in
rable to the binary orbit, providing since these sizes are below the canonical SS433
unprecedented spatial information on spatial resolution of an optical interferom-
their accretion-ejection mechanisms. eter such as the Very Large Telescope SS433 is unique in the Galaxy as the
In particular, observations of the hyper- Interferometer (VLTI), which is around 3 only known steady hypercritical accretor;
critical accretor SS433 have revealed a mas for a baseline of 100 metres. There- the donor star provides the compact
variety of spatial structures at the heart fore, in order to get to such scales, exqui- object (the nature of which remains enig-
of this exotic microquasar, including site precision in the interferometric observ- matic, but is likely to be a black hole) with
bipolar outflows, super-Keplerian equa- ables is required, which is best achieved matter at a rate hundreds of times above
torial outflows and extended baryonic with spectrally resolved measurements Eddington (see, for example, Fabrika,
jets photoionised by collimated ultravio- using strong emission lines. This tech- 2004 for a review of SS433). The resulting
let radiation. nique is called spectral differential inter- geometrically and optically thick super-
ferometry and it can be used to acquire critical accretion disc thermally down-
robust velocity-resolved microarcsecond grades the X-ray radiation produced
X-ray binaries (XRBs) are composed of a (µas) spatial information. close to the compact object (and typically
compact object (neutron star or black seen in ordinary X-ray binaries) to ultra
hole) accreting matter from its donor star. GRAVITY has led to a breakthrough in violet (UV) and optical wavelengths, turn-
The accretion process leads to a variety the ability to fringe-track on faint objects, ing the compact object into an accretion-
of inflow-outflow structures such as discs, allowing interferometric quantities to powered quasi-star that outshines its
streams, winds and jets. While large- be measured in the near-infrared (NIR) donor star at all wavelengths. In addition,
scale jets are often resolved with very at high spectral resolution (R ~ 4000) the enormous radiation pressure leads
long baseline interferometry (VLBI) at with unprecedented precision. When to powerful outflows producing strong
radio wavelengths, capable of achieving applied to X-ray binaries, this has led to emission lines, seen not only from the
approximately milliarcsecond (mas) spatial the first spatially resolved observations ~ 2000 km s –1 accretion disc winds (the
resolution, the inner parts of the accre- of accretion-ejection structures at NIR so-called “stationary” lines) but also
tion-ejection structures, at scales compa- wavelengths. Here, we review pioneering from the ~ 80 000 km s –1 (0.26c) highly
2.50
a) Brδjet1 Brγ Brγjet1 Brγjet2 Je
2.25 t in
He I He I jet1 He I jet2 N te n
2.00 sit
yp
rofi
F(λ)/Fcont
1.75
le
1.50
E -1
1.25
1.00
1.7
b)
ma
1.2 s
UT1–UT3 - 1/
Normalised visibility
1.0
0.8
c)
Normalised phases (degrees)
10 0.2
UT1–UT3 6c
0 1 mas
–10
> 15 mas
2.05 2.10 2.15 2.20 2.25 2.30
λ (μm)
Figure 1. Left: Spectrum (a), differential visibility ary” Brg line alternates between a bipolar- nation with spectroscopic observations,
amplitudes (b), and visibility phases (c) across the
outflow dominated mode (aligned with have shown that SS433 is UV-dominated
“stationary” and baryonic jet emission lines of
SS433 in the 2016 GRAVITY observation. The best- the jets) to an equatorial-outflow domi- even in the jet funnel. This is important
fit exponential model for the jet emission (illustrated nated mode (perpendicular to the jets) in the context of the acceleration mecha-
in the schematic on the right) is shown in black. (Figure 2). Although the presence of nism of the ~ 0.26c jets by line-locking
equatorial outflows in SS433 had been (Milgrom, 1979), which requires intense
collimated baryonic jets, with the latter’s well established from radio observations collimated radiation, as well as in the
precession creating its idiosyncratic (for example, Blundell et al., 2001), this possible relation between SS433 and
moving lines across the X-ray and optical is the first time that velocity and size ultraluminous X-ray sources (ULXs).
spectra (Margon et al., 1979). could be combined to show that the out- Future observations of several hours in a
flows are super-Keplerian. The outflows single night could directly detect the
These unique properties — a very bright support the conclusion that both the relativistic motion (8 mas d –1) of the bary-
accretion disc at optical wavelengths compact object and the donor star in onic jets, providing a further probe of
and strong, broad and variable emission SS433 overfill their Roche lobes signifi- their heating mechanism and an accu-
lines — make SS433 the ideal XRB for cantly, losing mass through their outer rate, self-consistent distance to SS433.
NIR interferometry, also providing the Lagrangian points, and that the transfer
only opportunity to spatially study a of specific angular momentum between
supercritical accretion disc and its out- the binary and the disc-like outflows is Spatially resolved wind accretion in
flows. GRAVITY observations of SS433 very significant for the binary evolution. BP Cru
carried out in 2016 and 2017 (GRAVITY Future observations of several hours in a
Collaboration et al., 2017b; Waisberg et single night could harvest the full power BP Cru (GX 301-2) is composed of an
al., 2019a,b) have revealed a marginally of aperture synthesis and provide velocity- X-ray pulsar accreting from the wind of its
resolved NIR continuum consisting of (i) resolved, model-independent images hypergiant B1Ia+ companion (Kaper et
the central unresolved binary (with a size of the complex outflow structure beyond al., 2006). The latter has unusually power-
< 0.5 astronomical units [au]), and (ii) simple geometric models. ful winds (~ 10 – 5 M⊙ yr –1) for a donor star
extended emission of size ~ 40 au in the in an XRB, which lead to strong emission
form of a wind and/or disc (contributing The GRAVITY observations have also lines of HeI and Brg from its extended
~ 20% of the K-band flux). Much more spatially resolved the optical jets of wind in its K-band spectrum. In addition,
information, however, is gathered from SS433 for the first time, revealing expo- its unusually high eccentricity (e = 0.46)
the spectrally resolved differential visibility nential profiles that extend to over 20 au makes it an ideal target for probing the
amplitudes and phases across the many (i.e., several tens of times the binary size) influence of the gravitational and radiation
emission lines (Figure 1). and which peak surprisingly close to the fields of the pulsar on the surrounding
central binary (Figure 1). These observa- circumstellar environment (for example,
For instance, the observations have tions suggest that optical jets are heated Blondin, 1994).
shown that the double-peaked “station- by collimated UV radiation and, in combi-
0.8 0.8
17 July 2016 9 July 2017
Bipolar outflow Equatorial Outflow
0.6 0.6
0.4 0.4
Brγ stationary
Declination (mas)
Declination (mas)
Approaching jet
– 0.2 – 0.2 Receding jet
– 0.6 – 0.6
1.00 0.75 0.50 0.25 0.00 – 0.25 – 0.50 – 0.75 –1.00 1.00 0.75 0.50 0.25 0.00 – 0.25 – 0.50 – 0.75 –1.00
Right ascension (mas) Right ascension (mas)
Figure 2. Velocity-resolved emission centroids of the stellar wind facing the compact Acknowledgements
across the double-peaked Brg “stationary” line for
object. In addition, asymmetries revealed
observations in 2016 (left) and 2017 (right). The See page 23.
e mission centroid of the spatially resolved baryonic by the differential visibility phases across
jets is also shown. The black circles correspond to the emission lines may point to an addi-
the estimated binary orbit size. tional component, possibly a stream of References
enhanced density which has been pos-
Blondin, J. M. 1994, ApJ, 435, 756
The spectral differential visibilities meas- ited to exist in the system from the analy- Blundell, K. et al. 2001, ApJ, 562, L79
ured by GRAVITY (GRAVITY Collabora- sis of X-ray light curves (Leahy & Kotska, Fabrika, S. 2004, Space Science Reviews, 12, 1
tion et al., 2017a) reveal an extended wind 2008). Further observations at different GRAVITY Collaboration et al. 2017a, ApJ, 844, 177
GRAVITY Collaboration et al. 2017b, A&A, 602, L11
with a size several times the stellar radius, orbital phases could take advantage of
Kaper, L. et al. 2006, A&A, 457, 595
which is also significantly distorted — the significant eccentricity in order to Leahy, D. & Kostka, M. 2008, MNRAS, 384, 747
being more extended on the side that is disentangle intrinsic variability of the wind Margon, B. et al. 1979, ApJ, 233, L63
shielded from the pulsar — and which from the distortion caused by the pulsar Milgrom, M. 1979, A&A, 78, L9
Waisberg, I. et al. 2019a, A&A, 623, A47
could be caused by the X-ray ionisation accretion.
Waisberg, I. et al. 2019b, A&A, 624, A127
DOI: 10.18727/0722-6691/5170
GRAVITY Collaboration (see page 20) massive binary η Car with GRAVITY from the detection of planets to mapping
across two spectral lines: He I and Brg. the cores of active galactic nuclei (AGN).
514.0 km s –1 551.0 km s –1 3
Norm. flux
10
2
0
Wavelength (m)
–10 1
2.162 2.164 2.166 2.168 2.170
1e–6
10 0 –10
Right ascension (mas)
textures (like granular surfaces, or point- These algorithms minimise (i) the differ- The massive binary at the core of h Car
like objects) while the lowest ones trace ence between the data and the visibilities
extended textures (like edges and con- obtained from the model image (i.e., Located at the core of the “Homunculus
tours). An image is, therefore, composed the likelihood term), and (ii) the value of Nebula” (see Figure 1) at a distance of
of an infinite number of frequencies. one or several priors (i.e., the regularis- 2.3 kiloparsecs, h Car is a very massive
However, interferometers only sample a ers), which are defined based on the and intriguing object. Indirect observations
few of them. knowledge of the source (see Sanchez- suggest that a binary with a period of
Bermudez et al., 2018). Reconstruction 5.54 years resides in its core. The primary,
Thus, recovering an image from interfero- packages available to the community hA, is supposedly a star with a mass of
metric data is an “ill-posed” problem with optimise through gradient-descent (for around 100 M⊙, while the secondary, hB,
more unknowns (pixels in the image) than example, MiRA: Thiébaut, 2008; BSMEM: appears to be a hotter star, perhaps a
constraints (data). Reconstructing images Buscher, 1994) or Monte-Carlo Markov- giant O-star, with a mass of around 30 M⊙,
requires the use of iterative regularised Chain methods (for example, SQUEEZE: but around 100 times fainter than the pri-
least-squares minimisation algorithms. Baron & Kloppenborg, 2010). mary. Different observations suggest that
hA exhibits a very dense and slow wind was therefore selected as a target for ionised He I is formed from (i) a portion of
that shocks with a much faster and the Guaranteed Time Observations (GTO) the primary wind, which is photoionised
lighter wind from the secondary. GRAVITY programme (GRAVITY Collabo- by the strong ultraviolet radiation of the
ration, 2017), with the objective of carry- hB wind, and (ii) by the shocked material
The existence of hB produces several ing out a long-term (monitoring) analysis in the cavity walls. To properly quantify
changes in the morphology of hA’s wind. of the wind-wind collision cavity through this scenario new spectro-interferometric
In particular, it photoionises part of the interferometric imaging. images are required in combination with
primary wind, changing the strength of dedicated modelling. Two additional
lines such as Ha, He I, Fe II, or Ne II The first reconstructed images, pre- imaging epochs in 2018 and 2019 have
(Mehner et al., 2010, 2012; Madura et al., sented in GRAVITY Collaboration (2018), been obtained with GRAVITY. From the
2012). 2D radiative transfer models and included data taken during the commis- preliminary analysis of data taken in 2018,
3D hydrodynamical simulations of the sioning phase (in 2016) of GRAVITY we can confirm that the morphology of
wind-wind collision scenario suggest that and through regular P100 programmes the wind-wind collision zone changes
the high-velocity secondary wind pene- (in 2017). h Car was observed with depending on the orbital phase of the
trates the slow and dense primary wind the Auxiliary Telescopes (ATs) using the secondary (Figure 2). As demonstrated in
creating a low-density cavity in it, with high-spectral-resolution mode of the case of h Car, GRAVITY spectro-in-
thin and dense walls where the two winds GRAVITY. This setup allowed us to terferometric imaging provides unique
interact (Madura et al., 2013; Clementel resolve several spectral lines across the information that can help to characterise
et al., 2015a,b). target’s spectrum and thereby to monitor the physics associated with the morphol-
the morphologies of the core at different ogy of complex systems at the highest
Several attempts have been made to Doppler velocities. In particular, we angular resolution currently possible in
map the core of h Car and to peer into focused our efforts on mapping Brg and the near-infrared.
the structure of the wind-wind collision the He I 2s–2p lines.
region, and of the binary itself, at scales
of 5–10 astronomical units (au) or Images were recovered using SQUEEZE. References
2–4 mas. Long-baseline infrared interfer- Prior information necessary for the Baron, F. & Kloppenborg, B. 2010, Proc. SPIE, 7734,
ometry has been a d ecisive technique reconstruction was included in both the 77344D
for such studies (van Boekel et al., 2003; spatial and spectral domains to obtain Buscher, D. F. 1994, Very High Angular Resolution
Weigelt et al., 2007). Astronomical simultaneous images of 35 different Imaging, IAU Symposium, 158, 91
Clementel, N. et al. 2015a, MNRAS, 450, 1388
Multi-BEam combineR (AMBER) obser spectral channels, with a resolution as Clementel, N. et al. 2015b, MNRAS, 447, 2445
vations in 2014 allowed, for the first time, good as 1.75 mas (4 au; see Figure 1). GRAVITY Collaboration 2017, A&A, 602, A94
the recovery of aperture-synthesis Compared with the 2014 AMBER images, GRAVITY Collaboration 2018, A&A, 618, 125
images, at a resolution of ~ 6 mas of the the GRAVITY Brg ones revealed struc- Mehner, A. et al. 2010, ApJ, 710, 729
Mehner, A. et al. 2012, ApJ, 751, 73
wind-wind collision cavity across Brg tural changes associated with the orbital Madura, T. I. et al. 2012, ApJ, 647, L18
(Weigelt et al., 2016). motion of the secondary. In particular, Madura, T. I. et al. 2013, MNRAS, 436, 3820
a bright “clump” is observed towards the Sanchez-Bermudez, J. et al. 2018, Experimental
southeast of the central core, which was Astronomy, 46, 457
Thiébaut, E. 2008, Proc. SPIE, 7013, 701311
Observing h Car with GRAVITY identified as part of shocked wind flowing van Boekel, R. et al. 2003, A&A, 410, L37
along the inner cavity walls after the last Weigelt, G. et al. 2007, A&A, 464, 87
The unique characteristics of h Car make hB periastron in 2014. Weigelt, G. et al. 2016, A&A, 594, A106
it a good candidate for increasing our
understanding of the role of multiplicity in The He I images revealed, for the first
shaping the fate of stars at the upper time, the distribution of this element in
end of the Initial Mass Function (IMF). It h Car’s core. We suggest that the partially
Markus Wittkowski 1 can form in their atmospheres, and are and Alfvén waves (for example, Airapetian
Sara Bladh 2 subsequently expelled into the interstellar et al., 2010; Cranmer & Saar, 2011;
Andrea Chiavassa 3 medium via stellar winds. Yasuda & Kozasa, 2019; Rau et al., 2019).
Willem-Jan de Wit 1 Radiative pressure is currently being
Kjell Eriksson 2 Both AGB stars and RSGs are affected implemented in global CO5BOLD
Bernd Freytag 2 by pulsation and convection, but RSGs models. Magneto-hydrodynamical effects
Xavier Haubois1 show lower variability amplitudes than can, in principle, be described by
Susanne Höfner 2 AGB stars. For AGB stars, it has been CO5BOLD models (Freytag et al., 2012;
Kateryna Kravchenko 1 shown that pulsation and convection lead Steiner et al., 2014), but an application
Claudia Paladini 1 to strongly extended molecular atmos- to AGB and RSG stars requires further
Thibaut Paumard 4 pheres, where the temperature is low work.
Gioia Rau 5, 6 enough for dust condensation. Radiation
Peter R. Wood 7 pressure on dust then gives rise to a
general mass outflow as the surrounding Pilot study with GRAVITY
gas is dragged along through friction
1
ESO (for example, Höfner & Olofsson, 2018). Time-series of interferometric observa-
2
Uppsala University, Department of tions provide the strongest tests of
Physics and Astronomy, Sweden For RSGs, it has been speculated that dynamical processes in the atmospheres
3
Université Côte d’Azur, Laboratoire the same processes may explain their of evolved stars, as they spatially resolve
Lagrange, Nice, France mass loss. However, Arroyo-Torres et al. the star and provide constraints on differ-
4
LESIA, Observatoire de Paris, Meudon, (2015) showed that current dynamic ent atmospheric layers, following the
France model atmospheres of RSGs, based on variability cycle of the star. However, such
5
NASA, Goddard Space Flight Center, pulsation and convection alone, cannot time-series are still very rare.
Greenbelt, USA explain the observed extensions of RSG
6
Catholic University of America, Depart- atmospheres, or how they can reach dis- Wittkowski et al. (2018) recently con-
ment of Physics, Washington, DC, USA tances where dust can form. This points ducted a pilot study measuring the varia-
7
Research School of Astronomy and to missing physical processes in current bility of the continuum radius and of
Astrophysics, ANU, Canberra, Australia RSG dynamic models. It translates into extended molecular layers for the oxy-
uncertainties in our general understand- gen-rich Mira star R Peg during science
ing of mass loss, as such processes may verification and early (P98) science
Mass loss from cool evolved stars is an to some degree also affect the atmos- operations, using the newly available
important ingredient of the cosmic mat- pheric structures of AGB stars and other near-infrared K-band beam combiner
ter cycle, enriching the Universe with cool giants. GRAVITY (GRAVITY Collaboration, 2017)
newly formed elements and dust. How- at the VLTI. This became possible
ever, physical processes that are not because of the improved performance of
considered in current models represent 1D and 3D model atmospheres the GRAVITY instrument compared to, for
uncertainties in our general under- example, the Astronomical Multi-BEam
standing of mass loss. Time-series of Significant advances are being made in combineR (AMBER), with an increased
interferometric data provide the strong- the development of dynamic atmosphere precision in visibilities, data for six base-
est tests of dynamical processes in the models of cool evolved stars. Latest lines in one snapshot, and a spectral
atmospheres of these stars. Here, we developments include 1D DARWIN (Bladh resolution of about 4000 across the full
present a pilot study of such measure- et al., 2019), and 3D CO5BOLD radiative K-band.
ments obtained with the GRAVITY hydrodynamics (RHD) simulations
instrument on the Very Large Telescope (Freytag et al., 2017; Höfner & Freytag, We showed that the continuum size and
Interferometer. 2019). In contrast to existing CO5BOLD the size in a bandpass that is dominated
and CODEX models, DARWIN models by water vapour were anti-correlated
include the wind acceleration region, with the visual light-curve. The size in the
Cool evolved stars which affects atmospheric structure and CO (2–0) line instead follows the visual
molecular features (Bladh et al., 2013, light-curve more closely, indicating a dif-
Asymptotic giant branch (AGB) and red 2015; Höfner et al., 2016), and may ferent — possibly more stable — behav-
supergiant (RSG) stars are located in account for some of the previously found iour of CO compared to water vapour
the Hertzsprung–Russell diagram at low discrepancies between AGB star models (Figure 1). The wavelength-dependent vis-
effective temperatures (about 2500– and interferometric observations. Addi- ibility variations could be reproduced by a
4500 K). They are major contributors to tional processes that may contribute set of CODEX (Ireland et al., 2008, 2011)
the integral luminosity of stellar systems, to larger atmospheric extension in RSG dynamic model atmospheres at phases
and they are major sources of the chemi- dynamic models include radiation pres- between 0.3 and 0.6. However, we
cal enrichment of galaxies. Owing to the sure on molecular lines (Josselin & Plez, noticed the following issues: (1) best-fit
low temperatures, molecules and dust 2007) or the effects of magnetic fields model phases did not correspond well
12 8
14
7600 7700 7800
Time (JD-2 450 000)
with observed phases, and (2) the phere (Kravchenko et al., 2018, 2019). Ireland, M. et al. 2011, MNRAS, 418, 114
Kravchenko, K. et al. 2018, A&A, 610, A29
observed amplitude of the continuum Combined with spectro-interferometric
Kravchenko, K. et al. 2019, A&A, 632, A28
radius is 14% — this is smaller than GRAVITY observations on the VLTI, Rau, G. et al. 2019, ApJ, 882, 37
predicted by CODEX model atmospheres the tomographic method will permit a Steiner, O. et al. 2014, PASJ, 66, S5
(45%–67%), and closer to those pre- simultaneous spectral and spatial char- Josselin, E. & Plez, B. 2007, A&A, 469, 671
Wittkowski, M. et al. 2018, A&A, 613, L7
dicted by 3D RHD simulations (Freytag acterisation of AGB and RSG star atmos-
Yasuda, Y. et al. 2019, ApJ, 879, 77
et al., 2017). The data covered only four pheres. By extracting interferometric
epochs, and more are needed to be visibilities at wavelengths contributing to
meaningfully compared to 3D models, different masks, we can measure the cor- Links
which show strong intra-cycle and cycle- responding geometrical extents of the 1
RAVITY Science Verification: https://www.eso.
G
to-cycle irregularities. atmosphere and recover the link between org/sci/activities/vltsv/gravitysv.html
optical and geometrical depth scales. 2
GRAVITY consortium: http://www.mpe.mpg.de/ir/
gravity
Outlook
Acknowledgements
We plan to extend the GRAVITY pilot Based on observations made with the VLT Interfer-
study described above to a larger sample ometer at Paranal Observatory. We thank the
of cool evolved stars, and in particular GRAVITY Science Verification team1, the GRAVITY
to include a comparison of AGB stars, for consortium 2, the GRAVITY Collaboration (see
page 20), and the ESO science operation team for
which current models successfully pre- the development and operations of GRAVITY, and
dict observed extensions, and RSG stars, for their great support.
for which models and observations show
strong discrepancies in this respect. We
References
need a denser and wider phase sampling
compared to our plot study, including Airapetian, V. et al. 2010, ApJ, 723, 1210
intra-cycle and cycle-to-cycle variations, Arroyo-Torres, B. et al. 2015, A&A, 575, A50
to be able to make meaningful compari- Bladh, S. et al. 2013, A&A, 553, A20
Bladh, S. et al. 2015, A&A, 575, A105
sons to the latest dynamic models. Bladh, S. et al. 2019, A&A, 626, A100
Cranmer, S. R. & Saar, S. H. 2011, ApJ, 741, 54
We will be able to use more, and better- Freytag, B. et al. 2012, JCoPh, 231, 919
defined, atmospheric layers compared to Freytag, B. et al. 2017, A&A, 600, A137
GRAVITY Collaboration 2017, A&A, 602, A94
our pilot study by applying a tomographic Höfner, S. et al. 2016, A&A, 594, A108
method that relies on spectral masks Höfner, S. & Olofsson, H. 2018, A&ARv, 26, 1
selecting lines that form in given ranges Höfner, S. & Freytag, B. 2019, A&A, 623, A158
of optical depths in the stellar atmos- Ireland, M. et al. 2008, MNRAS, 391, 1994
GRAVITY Collaboration (see page 20) High angular resolution observations are and a semi-major axis of 18.2 ± 0.3 au.
crucial to pinning down the dominant Additionally, we determined a new Orbit
mode of massive star formation. One of for q1 Ori D2, with a semi-major axis of
GRAVITY observations reveal that most the closest massive star forming regions 0.77 ± 0.03 au and a period of 53.05 ±
massive stars in the Orion Trapezium is the Orion Nebula Cluster, located at a 0.06 days.
cluster live in multiple systems. Our distance of 414 ± 7 pc (for example,
deep, milliarcsecond-resolution interfer- Reid et al., 2014). As such the Orion Neb-
ometry fills the gap at 1–100 astrono ula has been the target of many previous Most massive stars live in multiple
mical units (au), which is not accessible observations. The superb angular resolu- systems
to traditional imaging and spectros- tion and sensitivity of GRAVITY using
copy, but is crucial to uncovering the the VLTI can reveal details on the crucial Massive stars are more often found in
mystery of high-mass star formation. scales of 1–100 au, which had remained multiple systems than are lower mass
The new observations find a signifi- mostly unexplored until now. stars. For example, Duchene & Kraus
cantly higher companion fraction than (2013) and Sana et al. (2014) found
earlier studies of mostly OB associa- increasing numbers of stars in compan-
tions. The observed distribution of Observations with GRAVITY ion systems with higher stellar mass.
mass ratios declines steeply with mass Additionally, the average number of com-
and follows a Salpeter power-law We observed the 16 brightest, most mas- panion stars increases with higher mass.
initial mass function. The observations sive stars in the Orion Nebula Cluster, Our observations confirm this trend
therefore exclude stellar mergers as with masses between 2 and 44 M⊙. The and our results are comparable to those
the dominant formation mechanism for observations were mostly done with of Sana et al. (2014). Orion’s O-type stars
massive stars in Orion. the Auxiliary Telescopes in astrometric have an average of 2.3 ± 0.3 companions.
configuration. Data were reduced with
the standard GRAVITY pipeline (GRAVITY Plotting the number of all our observed
The formation of massive stars Collaboration, 2017). The interferometric stars and their companions against stellar
data were then fitted to a binary star mass, we find the mass function well
The formation of massive stars remains a model, providing the flux ratio of the com- described by a power law with an expo-
mystery. Hidden in their parental gas panion to the main star, and the separa- nent of G = 1.3 ± 0.3 (Figure 2). This
and dust clouds, it is unclear how their tion vector between the two components matches the initial mass function (IMF)
seeds can accrete so much matter before (GRAVITY Collaboration, 2018). for field stars (see, for example, Salpeter,
the repulsive forces from thermal pres- 1955).
sure and radiation prevent the formation We focused first on the central region,
of a protostar. The most discussed the Orion Trapezium Cluster, home of Ori- To constrain star formation scenarios,
scenarios are competitive accretion and on’s most massive, visible star, q1 Ori C. we compare predictions to our observa-
core accretion (see, for example, Tan et The 16 observed objects have a total tions. For both core accretion and com-
al., 2014 and references therein). Another of 22 companions; see Figure 1 for an petitive accretion, the number of stars
possibility is the collision of two stars, overview. in companion systems and the number of
merging into a more massive star. companions should rise with mass
With GRAVITY, we found three previ- (Clarke, 2001). Therefore, both scenarios
Core accretion is a scaled-up version of ously unknown companions and we con- would match our observations. The situa-
standard star formation applicable to firm a suspected companion for n Ori tion is different for the correlation
stars similar to our Sun. In this scenario it (Grellmann et al., 2013). The newly dis- between the companion masses. While
is a single core that accretes its mass covered stars belong to the systems of competitive accretion shows no clear
independently of other sibling cores. The q1 Ori B, q2 Ori B, and q2 Ori C. We deter- correlation between the primary and sec-
mass of the star is then set at the begin- mined their separation, and from the ondary mass, with a mass distribution
ning of the process, determined by the flux ratio we could estimate the masses that could follow a Salpeter IMF (for
available mass in the accretion volume. of all new companions (see GRAVITY example, Tan et al., 2014) or a top-heavy
An alternative explanation is formation by Collaboration, 2018 for more details). companion mass distribution (Bate,
competitive accretion (for example, Tan q1 Ori B is a system of particular interest, Bonnell & Bromm, 2002), core accretion
et al., 2014), where several cores com- as it consists of six objects in total. These results in a strong correlation between
pete for the available mass, culminating in objects are all gravitationally bound, the companion masses, which we do not
hierarchical systems with stars of differ- though it is suspected that the system is observe. Also, the companion separation
ent masses. Unlike in the core-accretion only temporarily stable (Close et al., should correlate with system mass for
model, their masses are not pre-defined, 2013). q1 Ori C is accompanied by two core accretion. For competitive accretion,
but depend on the interaction with each companion stars, one spectroscopic the separation should inversely correlate
other. A third possibility for the formation companion and one known companion with system mass (Bonnell & Bate, 2005).
of massive stars is stellar mergers, where with a determined orbit. With GRAVITY In Orion, we observe no correlation
two colliding stars end up in a more mas- observations, we could refine the orbit of between separation and system mass
sive object. q1 Ori C2 to have a period of 11.4 ± 0.2 yr (Figure 3), which is inconsistent with
B
E 100 au
θ2 Ori C
D
A
θ2 Ori B C
40 au 580 au
F 0.7 au
10 au
observed
IMF, Γ = 1.3
θ2 Ori A (39.0 ± 14.0 M )
IMF, Γ = 1.6
10 –1 IMF, Γ = 1.0 θ1 Ori C (33.0 ± 5.0 M )
Number of stars
Nu Ori (16.01 ± 3 M )
System & primary mass (M )
θ2 Ori C (4 ± 1.0 M )
101
Mass (M )
θ Ori E (2.81 ± 0.05 M )
1
either scenario. In addition, competitive stars live in multiple systems. We do not Acknowledgements
accretion predicts an anti-correlation see a strong preference for either core
See page 23.
between the mass ratio of the companion collapse or competitive accretion among
to primary star and their separation, the massive stars of Orion. The Salpeter
which we do not see in our data. If stellar IMF hints towards competitive accretion, References
collisions were the dominant formation whereas the lack of correlations between
Bate, M. R., Bonnell, I. A. & Bromm, V. 2002,
process, we would expect a strong devi- separation, system mass, primary and MNRAS, 336, 705
ation from the Salpeter IMF (Moeckel & companion masses contradicts it. We Bonnell, I. A. & Bate, M. R. 2005, MNRAS, 362, 915
Clarke, 2011). Thus we can exclude stellar can exclude the collision of stars as the Clarke, C. J. 2001, The Formation of Binary Stars,
IAU Symposium, 200, 346
mergers as the dominant formation main mechanism for the formation of high
Close, L. M. et al. 2013, ApJ, 774, 13
mechanism for massive stars in Orion. mass stars in Orion, which would result Duchêne, G. & Kraus, A. 2013, ARA&A, 51, 269
in a strong deviation from the Salpeter GRAVITY Collaboration 2017, A&A, 602, A94
IMF. Our GRAVITY results highlight the GRAVITY Collaboration 2018, A&A, 620, A116
Grellmann, R. et al. 2013, A&A, 550, 531
Summary & conclusions crucial role of interferometry in filling the
Kroupa, P. 2001, MNRAS, 322, 231
gap between 1 and 100 au, which is Moeckel, N. & Clarke, C. J. 2011, MNRAS, 410, 2799
We probed the Orion Nebula for massive not accessible with traditional imaging Reid, M. J. et al. 2014, ApJ, 783, 130
multiple star systems with separations and spectroscopic techniques. Salpeter, E. E. 1955, ApJ, 121, 161
Tan, J. C. et al. 2014, Protostars and Planets VI, ed.
between 1 and 100 au. Almost all massive
Beuther, H. et al., (Tucson: Univ. of Arizona), 149
DOI: 10.18727/0722-6691/5173
GRAVITY Collaboration (see page 20) are born in and/or migrate into the inner- gas might accrete onto the star through
most regions close to the host star. As magnetospheric accretion or be launched
discs evolve, different phenomena such through winds and jets, where dust is
More than 4000 exoplanets are known as photoevaporation, mass-loss through thermally processed, sublimated, and
to date in systems that differ greatly winds and jets, and dynamical clearing from where it can be redistributed into
from our Solar System. In particular, by newly-formed planets will disperse the the outer disc. Identifying dust traps
inner exoplanets tend to follow orbits disc material. Thus disc evolution and and other planetary signposts such as
around their parent star that are planet formation are linked processes. dynamical perturbations in the disc is
much more compact than that of Earth. Observing the inner regions with suffi- an important goal if we are to constrain
These systems are also extremely cient angular resolution is crucial for bet- inner planet formation mechanisms.
diverse, covering a range of intrinsic ter understanding the key physical pro-
properties. Studying the main physi- cesses at play and how they combine to
cal processes at play in the innermost lead to the formation of an exoplanetary The diverse nature of the inner discs
regions of the protoplanetary discs system.
is crucial to understanding how these In this contribution, we highlight GRAVITY
planets form and migrate so close to Thanks to high angular resolution imaging observations that reveal the morphology
their host. With GRAVITY, we focused in the optical range with the Spectro- of the inner dusty discs. The near infrared
on the study of near-infrared emission Polarimetric High-contrast Exoplanet emission detected with GRAVITY 1 and
of a sample of young intermediate- REsearch instrument (SPHERE; Beuzit et the Precision Integrated Optics Near-
mass stars, the Herbig Ae/Be stars. al., 2019), and at (sub-)millimetre wave- infrared Imaging ExpeRiment (PIONIER 2;
lengths with the Atacama Large Millime- Le Bouquin et al., 2011) arises mostly in
ter/submillimeter Array (ALMA partnership the dust sublimation front of the inner part
Dust in the innermost regions of the et al., 2015), rings, gaps, spiral arms, of the protoplanetary disc. We observe
young intermediate-mass stars warps, and shadows have been revealed wedge-shaped rims, with a smooth radial
in the outer disc on scales ranging from distribution of dust that is much wider
The formation and evolution of proto a few tens to a few hundreds of astro- than would be expected for a single dust
planetary discs are important stages in nomical units (au). GRAVITY uniquely component (GRAVITY Collaboration et
the lifetimes of stars. Terrestrial planets probes the innermost few au where hot al., 2019). We suggest that these inner-
Ae/Be mass range and need additional emission we measured with GRAVITY is GRAVITY Collaboration et al. 2019, A&A, 632, A53
Le Bouquin, J.-B. et al. 2011, A&A, 935, A67
observations. located at positions smaller than the
Lopez, B. et al. 2018, SPIE, 10701, 107010Z
critical radius where the gap is expected Maaskant, K. M. et al. 2013, A&A, 555, A64
to form as a result of to extreme-/far- Meeus, G. et al. 2001, A&A, 365, 476
Gap formation scenarios ultraviolet/X-ray heating, the discs in our Menu, J. et al. 2015, A&A, 581, A107
Vioque, M. et al. 2018, A&A, 620, A128
sample might be shaped by forming
Zhang, S. et al. 2018, ApJ, 869, L47
Gaps in protoplanetary discs can be young planets rather than by depletion
found in concentric arrangements from resulting from photoevaporation (Figure 2).
the inner regions out to large distances Notes
— as nicely evidenced by ALMA images With PIONIER, GRAVITY and MATISSE, 1
RAVITY operates in the near-infrared K-band, i.e.,
G
(Zhang et al., 2018). Clearing by dynami- the VLTI is perfectly equipped to reveal with wavelengths between 2 and 2.5 µm.
cal effects due to newly-born planets the gas and dust distributions in proto 2
PIONIER at the VLTI operates in the near-infrared
and photoevaporation by extreme- planetary discs at unprecedented angular H-band, i.e., with wavelengths between 1.5 and
1.8 µm.
and far-ultraviolet (EUV/FUV) and X-ray and spectral resolution. 3
M ATISSE at the VLTI operates in the mid-infrared
radiation from the central star are key L-, M-, and N-bands, i.e., with wavelengths
processes of disc dispersal through gap between 3 and 13 µm (Lopez et al., 2018).
and inner cavity formation. In the photo- References
evaporation scenario, gap formation ALMA Partnership et al. 2015, ApJ, 808, L1
takes a few 10 6 years and inner disc Beuzit, J.-L. et al. 2019, A&A, 631, A155
depletion takes about 10 5 years (Gorti et Gorti, U. et al. 2009, ApJ, 705, 1237
al., 2009). Since almost all the K-band Dullemond, C. & Dominik, C. 2004, A&A, 421, 1075
DOI: 10.18727/0722-6691/5174
GRAVITY Collaboration (see page 20) although there are many studies of the emission at 2.3 microns is a good tracer
outer part of the disc, there are very few of the hot inner gaseous disc. Therefore,
on the inner disc, in particular the inner spatially resolved observations of the CO
Near-infrared interferometry gives us gaseous disc. This hinders our under- ro-vibrational transitions are crucial to
the opportunity to spatially resolve standing of the physical processes taking constraining the dynamics and chemical
the circumstellar environment of young place in this inner part of the disc. In composition of the inner dust-free disc.
stars at sub-astronomical-unit (au) addition to the study of the size and
scales, which a standalone telescope shape of the continuum emission origi-
could not reach. In particular, the sensi- nating in protoplanetary discs around The source
tivity of GRAVITY on the VLTI allows young stellar objects (YSOs), GRAVITY’s
us to spatially resolve the CO overtone spectral resolution of up to R = 4000 51 Oph is a fast-rotating star (v sin i =
emission at 2.3 microns. In this article, facilitates studies of the gaseous compo- 267 km s –1; Dunkin, Barlow & Ryan, 1997)
we present a new method of using nent of circumstellar material. The two and is located at a distance of 123 par-
the model of the CO spectrum to recon- most prominent components are hydro- secs (Lindergren et al., 2016; GAIA Col-
struct the differential phase signal gen, in the form of the Brackett g and laboration et al., 2018). Its spectrum is full
and extract the geometry and size of higher levels of the Pfund recombination of atomic and molecular lines and it is
the emitting region. lines, and molecular gas as traced by CO one of the very few Herbig Ae/Be stars
ro-vibrational transitions. A direct tracer that shows bright 2.3-micron CO over-
of the gas in the disc is the CO molecule. tone emission, making it an ideal candi-
Protoplanetary discs at high angular The CO emission is present at different date for near-infrared interferometric
resolution scales throughout the disc: from the studies (Thi et al., 2005; Berthoud et al.,
outer, cooler regions detected at millime- 2007; Tatulli et al., 2008).
Circumstellar discs are crucial to under- tre wavelengths to the inner, warmer
standing how stars and planets form. regions detected at near-infrared wave- The CO spectrum of the star has been
They contain both gas and dust and, lengths. In particular, the CO ro-vibrational extensively studied spectroscopically
Figure 3. Cartoon of
a protoplanetary disc
based on Testi et al.
2400 K, GRAVITY, 1500 K, GRAVITY, SPHERE, NIR (2014). The different
boxes correspond to
CO bandheads NIR continuum scattered light different spatial scales
and temperatures of
the disc as observed by
different instruments.
0.1 au 1 au 10 au 100 au
ring, the angular extent of the CO, and ferometric observables can provide new References
the known distance to 51 Oph from GAIA insights into the geometry and size of the
Berthoud, M. G. et al. 2007, ApJ, 660, 461
Data Release 2, results in a direct meas- gaseous disc very close to the star. In Dunkin, S. K., Barlow, M. J. & Ryan, S. G. 1997,
urement of the mass of the central star of the case of 51 Oph we have been able, MNRAS, 286, 604
3.9 ± 0.6 M⊙ . for the first time, to observationally Gaia Collaboration et al. 2018, A&A, 616, A1
Lindegren, L. et al. 2016, A&A, 595, A4
constrain the physical properties of the
Muzerolle, J. et al. 2004, ApJ, 617, 406
GRAVITY has opened a new window gas at 0.1 au from the star; we find Tatulli, E. et al. 2008, A&A, 464, 55
enabling the use of molecular lines to physical properties consistent with those Testi, L. et al. 2014, Protostars and Planets VI,
probe the circumstellar environment of expected from LTE models of the ed. Beuther, H. et al., (Tucson: University of
A rizona Press), 339
young stars. This new technique of gas content of the disc (Muzerolle et al.,
Thi, W. F. et al. 2005, A&A, 430, L61
combining the spectrum fit with inter 2004).
ESO/P. Carrillo
Δ Declination (mas)
–5 0.00 0
5 PA = 54.2 BL = 37.1 PA = 70.85 BL = 120.0 PA = 84.45 BL = 88.8
– 50
0 – 0.02
–100
–5
2.164 2.166 2.168 2.164 2.166 2.168 2.164 2.166 2.168 – 0.04 –150
Wavelength (µm)
2
same technique could be applied in the u = 0.268 ± 0.03
future to measure the mass and dis- A + = 2.42
χ r2 = 1.20
tance of completely dark objects, such
a black hole. In fact, while black holes 1 E
Δ Declination (mas)
0 S
future by Gaia astrometry, gravitational uθ E = 0.51 mas
microlensing is the only known way to ξ = 0.20
find isolated black holes. Our detection E =
θ
angular Einstein radius at 2% precision: neous light curves from a space tele- initial discovery. And thanks to the excep-
1.87 ± 0.03 milliarcseconds (see Figure 1). scope in heliocentric orbit. The angular tional site conditions, we were able to
Interferometric resolution of images can Einstein radius measured from the observe it near the magnitude limit of
unleash microlensing’s unique potential VLTI resolution of microlensed images is VLTI/GRAVITY at a relatively high airmass
to find isolated stellar-mass black holes the other missing ingredient that can yield of ~ 1.5.
(BHs) lurking in the Galaxy by lifting the unambiguous determination of the lens
degeneracy between mass and distance mass and thereby definitively identify a The exceptional sensitivity of VLTI/
in the analysis of microlensing light BH lens. GRAVITY and the advent of all-sky bright
curves. transient surveys such as ASAS-SN
The lens of TCP J0507+2447 is not a BH and Gaia provide an unprecedented
LIGO/VIRGO’s astonishing discoveries of but a low-mass star. Nevertheless, it opportunity to obtain more resolved
merging BHs (Abbott et al., 2016) have can serve as a testbed for the above- microlensing images. We hope to carry
raised an important open question: how mentioned approach of lens mass deter- out a systematic survey towards the first
to form BHs with a few tens of solar mination. In an independent effort, definitive identification of an isolated stellar-
masses? Whether they are the end points another research team has measured the mass black hole.
of massive stars or have exotic origins lens flux with Keck adaptive optics
in the early universe, theories predict that images, and by combining this with our
isolated (single) BHs must exist. The precise VLTI angular Einstein radius References
relative frequency between the single and measurement, they find that the lens is a Abbott, B. et al. 2016, Phys. Rev. Lett., 116, 1102
binary BH populations can provide dwarf star of 0.58 ± 0.03 M⊙ (Fukui et al., Einstein, A. 1936, Science, 84, 506
crucial clues to the formation mechanism. 2019). Our team (Zang et al., 2019) has Dong, S. et al. 2019, ApJ, 871, 70
However, limited by the detection tech- measured the microlens parallax using Delplancke, F. et al. 2001, A&A, 375, 701
Fukui, A. et al. 2019, AJ, 158, 206
niques, all known stellar-mass BHs are the Spitzer light curves, and by combin- Gould, A. 2000, ApJ, 535, 928
found in binaries. ing the VLTI angular Einstein radius, the GRAVITY Collaboration 2017, A&A, 602, 94
lens is found to be 0.50 ± 0.06 M⊙ . The Nucita, A. et al. 2018, MNRAS, 476, 2962
Microlensing holds great promise in prob- good agreement between the results of Zang, W. et al. 2019, submitted to ApJ,
arXiv:1912.00038
ing the important yet uncharted parame- these two approaches demonstrates the
ter space of isolated BHs. Estimates robustness of our method. Remarkably,
by Gould (2000) suggest that, amongst around the peak of the light curve of
the microlensing events detected to date, TCP J0507+2447, there was a short-lived
many hundreds may involve BH lenses. anomaly lasting a few hours, suggesting
But thus far only a few BH candidates that the lens star has a 20-Earth-mass
have been reported. This is due to the planet at around 1 astronomical unit
mass-distance-velocity parameter degen- (Nucita et al., 2018; Fukui et al., 2019).
eracy, which makes it impossible to defin-
itively distinguish BHs from low-mass The possibility of using the VLTI to
stars. All existing BH candidates have rel- resolve microlensing images was first
atively long event timescales, which proposed by Delplancke et al. (2001), but
can be due to the large Einstein radii of it had proven to be extremely challenging,
BH lenses with high masses. But a large with numerous failed attempts prior
Einstein radius can also be produced to our observations. The major challenge
by a low-mass stellar lens at a close dis- had been the difficulty of identifying a
tance. Alternatively, a slow relative proper sufficiently bright target for the interfero-
motion between the lens and source metric observations. A confluence of
stars may induce a long timescale even lucky circumstances facilitated our suc-
with a moderate Einstein radius. cess. Unlike the vast majority of micro-
lensing events found by professional
To completely break the degeneracy, two wide-field surveys towards the Galactic
additional observables are required bulge, TCP J0507+2447 was serendipi-
besides the microlensing event timescale. tously discovered by the Japanese ama-
One is called the “microlensing parallax”, teur astronomer T. Kojima, and the
which depends on the Einstein radius source is at 800 pc towards the Galactic
projected onto the observer’s plane. It anti-centre, making it one of the closest
can be constrained for long events from microlensing events ever found. Our
the distortion of the light curves induced DDT proposal (2100.D-5031) was quickly
by the acceleration of the Earth while accepted, and an ongoing VLTI run
it orbits the Sun or by comparing the allowed our GRAVITY observations to be
ground-based observations with simulta- conducted within about a week of the
SPHERE photmetry
is obtained thanks to this exquisite post- 0.02
GRAVITY data
processing. Figure 1 shows GRAVITY
observations of the star AU Mic. The disc
0.01
of AU Mic has prominent structures, which
are resolved with SPHERE (Boccaletti,
Thalmann & Lagrange, 2015). GRAVITY 0.00
looks for point-like sources, of size
1.9 2.0 2.1 2.2 2.3 2.4 2.5
smaller than its interferometric resolution Wavelength (µm)
(< 3 mas). Larger objects are not seen in
the coherent flux of the interferometer. Figure 2. Upper panel: SPHERE/IRDIS image of CH4 absorption is detected. This gives
HR8799 acquired with a broadband H-filter (from
clues that help to characterise the atmos-
Wertz et al., 2017). As in Figure 1, we put the fringe-
The disadvantage of the single-mode tracker fibre on the star. The science spectrometer phere. The difficulty in interpreting the
interferometer is its field of view, which is fibre is on HR8799e. The sizes of the circles corre- data lies in the complex physical pro-
given by the diffraction limit of the tele- spond to the GRAVITY field of view. Below: GRAVITY cesses at work. Radiative transfer is used
K-band spectrum of HR8799e at spectral resolution
scope (60 mas in the K-band for the VLT to derive the pressure-temperature
500 (grey points) after 2 hours of integration. The
Unit Telescopes). Therefore, while the dashed curve is the K-band Gemini Planet Imager curves, but clouds at different altitudes,
fringe tracker fibre stays on the star, (GPI) spectrum from Greenbaum et al. (2018), show- with various compositions and possibly
the science fibre (which feeds the spec- ing speckle contamination. (GRAVITY Collaboration also heterogeneous, modify the tempera-
et al., 2019).
trograph) is placed at different positions ture distribution. Chemical disequilibrium
across the disc. This is how GRAVITY also adds complexity, with the necessity
hunts for exoplanets; it dithers the posi- GRAVITY as a way to characterise to add chemical timescales and mixing
tion of the fibre to cover a large area. In exoplanet atmospheres coefficients. In short, models need to
the case of AU Mic, we took advantage be challenged by observations, and
of the fact that we are only looking for a In addition to its dynamic range, GRA GRAVITY data is meeting that challenge.
planet along an edge-on disc, so we only VITY’s angular resolution yields i) precise
had to scan one line, thereby minimising astrometry (between 10 and 100 μas) In the near future, following the recent
the required telescope time. and ii) K-band spectra mostly unbiased upgrade of GRAVITY’s high-resolution
by s tellar light. Fortunately, such near- grism, a resolution of 4000 will be achiev-
In the resulting dataset, which covers only infrared spectra are rich in many molecu- able on exoplanets — a significant
the south-eastern part of the disc, no lar absorption lines: for example, H2O, increase compared to the previous reso-
detection was made. The dynamic range CO, CO2, CH4, N2O. We applied this to lution of 500 (because of limited sensitiv-
achieved by GRAVITY, with 15-minute HR8799e in GRAVITY Collaboration et al. ity). GRAVITY will therefore continue to
exposures, is 11 magnitudes at 120 mas (2019). HR8799e is the innermost object challenge models of exoplanetary atmos-
(5-s). At 250 mas the dynamic range is in a multi-planetary system. The angular pheres, requiring simulations with more
even higher, reaching 13.5 magnitudes. separation to its host star is 380 mas, resolution and more complex chemical
This is several magnitudes fainter than and the contrast is close to 11 magni- processes. One exciting prospect, for
what was achieved with aperture masking tudes in the K-band. The young planet example, is the detection of C13 isotopes
(Gauchet et al., 2016), and a completely has an effective temperature of 1150 K, (Mollière & Snellen, 2019). In parallel,
new domain compared to what could be still hot from its formation. The spectra, the recent development of atmospheric
done with ADI and SDI techniques on a shown in the bottom panel of Figure 2, parameter retrieval is an exciting new
single 8-metre telescope. show the CO absorption bands — no technique, which performs better than
ESO/M. Zamani
Students and project supervisors at the First ESO
Summer Research Programme pose for a photo
at the Supernova Planetarium and Visitor Centre
(see p. 57).
G.Hüdepohl (atacamaphoto.com)ESO
seen from above ESO’s
Petr Horálek 1
Paranal Observatory.
The VLT and VISTA
domes are coloured red
1
ESO during the sunset.
these rays can make for a dramatic through the atmosphere, the slower they In 1596, the ship carrying polar scientist
scene. They are also quite often visible become, and thus the more they are Willem Barents (c. 1550–1597) was
at sunset, when they shine over the tops refracted. The effect is measurable — even caught in the ice on what is now called
of clouds or through gaps within the when the Sun is high in the sky — and the Barents Sea at Novaya Zemlya Island.
clouds (Figure 3). They are called crepus- it can be important for navigators and There were two weeks until sunrise, but
cular rays and have been known by vari- astronomers. The extent of the refraction the tip of the solar disc had already peaked
ous poetic names in different cultures, varies from day to day, so the observed above the horizon (de Veer, 1876). Polar
for example, “Maui’s rope” — based on a sunset time can change from one day to explorer Sir Ernest Henry Shackleton
Maori tale; “Buddha’s rays” in parts of the next by as much as 5 minutes. We (1874–1922) also reported repeated
south-east Asia; and “Jacob’s ladder” cannot therefore rely on an ephemeris to sunrises “ahead of time” in the Arctic
in the UK (Lynch & Livingston, 2001). predict the observed sunset time with
perfect accuracy. Figure 4. During the early evening of 7 August 2017,
a partial lunar eclipse was visible in the sky above
When the Sun sets behind high moun- the ESO Headquarters. While the Moon was rising,
tains to the west, the mountains create A more specific variation of this phenom- significant anticrepuscular rays were visible in the
broad crepuscular rays by partially block- enon is visible in the Arctic regions. antisolar direction.
ing the light. Some dust or haze in the
air will serve to highlight the rays. Under
ESO/P. Horálek
perfect conditions, when crepuscular
rays can be followed all the way across
the sky, they appear to expand at first
and then narrow towards the east. In
reality, they remain in parallel but only
appear narrower at greater distances
from us. Crepuscular rays in the east are
called anti-crepuscular rays (Figure 4).
crescent Moon
the solar or lunar disc appears layered.
deformed by the atmos-
phere over the Pacific The subtle details of the layered lantern
Ocean, as seen from effect can best be observed with the
Paranal on 19 November Moon (see Figure 5), as the Sun often
2017. The effect of layers
saturates on pictures. The “lifting” effect
with different tempera-
ture and density in of the lower layers of the atmosphere
the atmosphere caused can also be seen on stars as they set
different parts of the (Figure 6).
Moon’s image to be
refracted by different
amounts as it neared Compared to other natural light phenom-
the horizon. ena, the green flash has an aura of mys-
tery and supernaturality. The green flash
(Shackleton, 1919). The explanation is Another effect of refraction is that the Sun is rare and very difficult to observe, but,
once again the atmosphere’s refraction of and Moon never appear perfectly round when the conditions are right, the observ-
light or rather, the variation in that refrac- as they approach the horizon, they er’s perseverance is rewarded with a
tion. In the case of Barents, the refraction appear flattened — as also evidenced in quick green gleam at the top of the set-
must have lifted the solar disc by about many of the photographs in this article. ting Sun, amid the red and orange shades.
5 degrees to peek over the horizon, The amount by which the rays are
but this is one of the more extreme refracted increases closer to the horizon, At sunset the blue and green light rays
cases. The phenomenon is known as the so rays from the upper edge of the solar are refracted a little more than the red
Novaya Zemlya effect (Lehn, 1979; Lehn & disc are refracted less than those from light, which means that the blue-green
German, 1981). the lower edge, flattening its shape. colours are “lifted” slightly more in the
Occasionally, the shape of the disc may sky. You could almost say there are two
be further disturbed if the layers in the solar discs — one blue-green and one
Figure 6. Several effects are seen on the setting
stars in this multi-exposure photo: atmospheric
atmosphere have different temperatures red (Young, 2013). Where they overlap,
scattering (reddening) and absorption, and differen- and refract the light by different amounts we see a yellowish-red Sun, with the red-
tial refraction (“lifting” of the stars). — called differential refraction. It gives dening of the Sun due to atmospheric
R. Wesson/ESO
P. Horálek/ESO
not so rare to see the Sun with a greenish
top, called the green segment, even before
it reaches the horizon (Figures 7 and 8).
ESO/S. Guisard
image. The ozone layer is responsible for
the spectral effect, since the light path
through the atmosphere is longer at
sunrise and sunset and the ozone layer
covers altitudes between about 12 and
40 km. The Chappuis band of ozone
absorbs light in a broad band centred
close to 590 nm in the orange. Owing to
the long path, the effect from this band
strengthens when the Sun is near the
horizon, effectively removing much of the
yellow, orange and red light. In the right
conditions this can produce the apparent
spectral separation needed to produce a
noticeable flash.
G. Blanchard/ESO
in rare circumstances, when the light
path is guided by differential refraction
to take an unusually long track through
the ozone layer, the flash is significantly
shifted towards shorter wavelengths to
produce the magnificent (and very rare,
Figure 9) blue flash; lucky is the person
who witnesses that!
ESO/C. Malin
caused by sunlight
whole mystery surrounding the flash
backscattering off tiny
and makes it something quite special to droplets of water in
look for. the atmosphere. Glories
appear at a point
directly opposite the
The green flash is best seen on a com-
position of the Sun,
pletely unobstructed western horizon, so they are only visible
like the view from the Paranal platform. at sunrise or sunset.
The weather must be very clear, and
the atmosphere needs to have complex
layers and a low aerosol content.
Happy hunting!
the horizon upwards as the Sun goes atmosphere that lie on the horizon or
down. The twilight arch is the arc of light high above our heads. The Earth shadow
The Earth shadow and the Belt of Venus that forms over the place where the Sun emerges and its visibility is best when
has set. It is usually red at the bottom, there is little dust or haze in the air.
As the Sun reaches the horizon, an yellow for a wide stretch above the red,
orange-yellow twilight arch forms to the and arches over in a peach-coloured, The clarity and low humidity of the air at
west, and the blue-grey Earth shadow green, turquoise or slightly purple band the high altitudes of the Atacama Desert
slowly rises to the east; stretching from that merges with the background colour. provide extraordinary opportunities to
It is created by scattered sunlight in the regularly observe the phenomenon called
Figure 10. The pink Belt of Venus and Earth shadow atmosphere. Even when the Sun has set, the Belt of Venus, followed by the projec-
as seen at ALMA. it can continue shining on parts of the tion of the dark Earth shadow onto the
D. Kordan/ESO
ESO/P. Horálek
Paranal Observatory on 25 January 2015. The bright
object is Venus. In this view rich dusk colours can
be seen. They were likely caused by volcanic ash
from the January 2015 eruption of Tongan volcano
and possibly even the 2014 eruptions of the Indone-
sian volcano Mount Sinabung.
The glory
est stars in the sky, can be visible in the 3
The telluric spectrum of the green flash (Fosbury, R.
2018): https://www.flickr.com/photos/
Observers at La Silla or Paranal can at sky very early after sunset (Figure 12) or
bob_81667/39604010580/
times see a phenomenon called a glory very shortly before sunrise. Usually just 4
Green and red rims (Young, A. T. 2013): https://aty.
when looking down on a cloud layer. minutes after sunset the brightest stars sdsu.edu/explain/simulations/std/rims.html
Glories are concentrated coloured rings appear while the Belt of Venus becomes
around the shadow of the observer’s larger and the sky above is “swallowed”
head (or the shadow of a camera; see by the Earth’s shadow. References
Figure 11). They occur by backscattering
Christensen, L. L. et al. 2016, The Messenger, 163, 40
on tiny water droplets in clouds or fog at Gladysheva, O. 2007, Solar System Research, 41, 314
the antisolar point and look like circular Acknowledgements Horálek, P. et al. 2016a, The Messenger, 163, 43
rainbows (Nussenzveig, 2011). This is a Horálek, P. et al. 2016b, The Messenger, 164, 45
The authors are grateful for helpful conversations Kundt, W. 2001, Current Science, 81, No. 4, 399
rather complicated case of Mie scattering with Bob Fosbury. Sarah Leach and Laura Hiscott Lee, R. L. 2015, Applied Optics, Vol. 54, Issue 4, B194
(and not the special divine importance are thanked for improvements to an earlier version Lehn, W. H. 1979, Journal of the Optical Society of
of a person!). If two people are standing of the text. America, Vol. 69, Issue 5, 776
on a mountain and look at both of their Lehn, W. H. & German, B. A. 1981, Applied Optics,
Vol. 20, No. 12, 2043
shadows, they will each see only one
Lock, J. A. 2015, Applied Optics, Vol. 54, Issue 4, B54
glory and claim it to be around their own Notes
Longo, G. 2007, Comet/Asteroid Impact and Human
head. Looking down at a plane’s shadow a
Society, 303
lease don’t forget that looking at the Sun itself,
P
when flying, you will often be able to see especially through an optical device (camera, tele-
Lynch, L. K. & Livingston, W. 2001, Color and Light in
Nature, (Cambridge: Cambridge University Press)
a glory on the cloud tops. scope, binoculars, etc.), is very dangerous, and Minnaert, M. G. J. 1993, Light and Color in the
could cause immediate blindness. Do not attempt Outdoor, (New York: Springer)
to observe the Sun unless you know what you are Moreno, H. et al. 1965, Science, 148, 364
doing.
Planets in dusk and dawn Nussenzveig, H. M. 2012, Scientific American, 306, 68
Shackleton, E. H. 1919, South: The Story of
Shackleton’s 1914–1917 Expedition, (London:
The clear air at the high altitudes of the Links W illiam Heinemann)
observatories makes dawn and twilight 1
de Veer, G. 1876, The Three Voyages of William
The colors of sunset and twilight (Corfidi, S. F.
colours very intense, but also followed by 2014, NOAA/NWS SPC): https://www.spc.noaa.
Barents to the Arctic Regions: 1594, 1595 and
1596, (London: Forgotten Books, 2017)
a steep darkening gradient. For these gov/publications/corfidi/sunset/ Waythomas, C. W. et al. 2010, Journal of
reasons, very bright objects, such as 2
Effect of atmospheric refraction on the times of Geophysical Research, 115, B12
planets in our Solar System or the bright- sunrise and sunset (Tong, Y. 2017, HKO): https:// Zieger, P. et al. 2013, Atmospheric Chemistry and
www.hko.gov.hk/m/article_e.htm?title=ele_00493 Physics, 13, 10609
Number of applicants
Anita Zanella 1 ESO Summer Research
Claudia Agliozzo1 Programme.
Richard I. Anderson 1
Fabrizio Arrigoni Battaia 2
Francesco Belfiore 1 50
Remco van der Burg 1
Chian-Chou Chen (T. C.) 1
Stefano Facchini 1
Jérémy Fensch 1
Prashin Jethwa 1 0
Rosita Kokotanekova 1
French
Dutch
Czech
Swedish
German
British
Danish
Spanish
Polish
Irish
Chilean
Finnish
Austrian
Belgian
Italian
Swiss
Portuguese
Federico Lelli 1
Anna Miotello 1
Anna Pala 1
Miguel Querejeta 1
Adam Rubin 1 Nationality
Dominika Wylezalek 1
Laura Watkins 1 as an opportunity that had been missing Programme overview
at Garching until now and decided to
organise a six-week-long Summer The programme started with a workshop,
1
ESO Research Programme at ESO for up to open to all ESO staff, on 1 July 2019.
2
Max Planck Institute for Astrophysics, seven university students. At this workshop the seven research pro-
Garching, Germany jects were introduced by the advisors,
A proposal was submitted by the and the students introduced themselves.
Garching ESO Fellows requesting funds An introduction to ESO was delivered
For the first time ever, a summer from the Director for Science to cover by the head of the ESO User Support
research programme was organised at travel costs and to provide a basic stipend Department Marina Rejkuba, and the
ESO Garching. Seven students, enrolled to cover lodging and living expenses for Director General greeted all participants
in universities all around the world, the students. The proposal was accepted from the control room of the La Silla
were selected from more than 300 and ESO Fellows, with the support of Observatory (he was visiting La Silla for
applicants. They each spent six weeks ESO administrative assistants, organised the total solar eclipse at the time).
from June to August 2019 carrying out the first-ever ESO Summer Research
a scientific project under the supervi- Programme. This involved booking ESO The students were each working on their
sion of teams of ESO Fellows and post- apartments and office space, setting up own research project, with the supervi-
docs, while engaging in the scientific the website 1, organising the application sion of one or more ESO Fellows, for the
life of ESO. The students carried out process and the selection of students, duration of the programme (Figures 2 & 3).
research in different fields of astron- planning and delivering a lecture series The schedule in the first three weeks also
omy, from comets to high-redshift gal- and, most importantly, designing and consisted of a set of eight lectures on
axies and from pulsating stars to leading the research projects. astronomical topics, a visit to the ESO
protoplanetary discs. In this report we Supernova including a planetarium show,
present the programme and describe The response from the community was and a telecon with Anita Zanella (an ESO
the main outcomes of the projects. incredible. More than 300 valid applica- Fellow observing at Paranal). The final
tions were received from university stu- three weeks were mainly focused on
dents in most Member States, from our the research projects, but with an addi-
Motivation and organisation Host Country Chile, and from ESO’s stra- tional two lectures and one visit to the
tegic partner, Australia (Figure 1). Partici- Extremely Large Telescope (ELT) primary
Summer studentship programmes for pants were selected by first distributing mirror test stand. Most of these additional
undergraduate students are becoming the applications amongst all potential activities and lectures were organised
the preferred way for an enterprising stu- supervisors for an initial ranking, followed following an explicit request from the stu-
dent to gain their first research experi- by a final selection by a committee com- dents who had expressed enthusiasm
ence; these programmes can last from a prising three fellows, one student and about the first set of lectures. Throughout
few weeks to months at top-class inter- one staff member. The final list included the duration of the programme the stu-
national universities or research centres. seven students — four females and three dents were among the most active
Such programmes have a wide range of males — from seven different countries. attendees of scientific activities at ESO
benefits to students and hosts alike. The After a short video interview all seven stu- Headquarters, including talks, science
ESO Fellows in Garching identified this dents accepted the offer. coffees, and informal meetings.
Figure 2. Student Tania Machado with her Figure 3. ESO Summer Research Programme stu-
supervisor Chris Harrison. dent Aisha Bachmann with supervisors Jeremy
Fensch and Remco van der Burg.
The last days were all focused on the This is a very challenging task because them during the next months and write
preparation of the most thrilling event for only the Hubble Space Telescope (HST) up her findings in a publication.
the students: their 15-minute presentation can spatially resolve UDGs at high red-
to be given in front of ESO staff, students shift, and cosmological surface bright-
and fellows during the final workshop ness dimming makes them extremely dif- Comet evolution from the Kuiper Belt
in the old ESO auditorium. This event ficult to detect. to a dormant comet in the near-Earth
was very well attended by ESO personnel asteroid population
(Figure 4) and showcased the great Aisha looked for UDGs, at redshifts Advisors: Rosita Kokotanekova
science that the students were able to beyond 1, in the deepest cluster images Student: Abbie Donaldson (UK & Ireland),
achieve during this relatively short pro- that were ever taken with the HST. She University of St Andrews, UK
gramme; some examples are described wrote a detection algorithm and tested it
in the next section. on mock galaxies that she inserted into This project focused on analysing photo-
the data; she then used the algorithm to metric observations of the comet 169P/
search for real UDG candidates. Finally, NEAT taken between February and June
Students and their research projects Aisha identified which UDGs, among the 2019 with the FOcal Reducer/low disper-
candidates she found, are cluster mem- sion Spectrograph 2 (FORS2) on the Very
Understanding the formation mechanism bers rather than projections along the line Large Telescope (VLT) and with the Wide
of galaxies at their extremes of sight by statistically comparing her Field Camera (WFC) on the Isaac Newton
Advisors: R
emco van der Burg & Jérémy detections with those of a reference field. Telescope (INT) on La Palma. Since the
Fensch Her preliminary results look extremely comet was observed close to aphelion
Student: Aisha Bachmann (German), interesting and Aisha aims to finalise and was therefore inactive, the photomet-
University Bochum, Germany
Figure 2. Conference photo. Zdenka Kuncic presented a special, and As already indicated, there were also
quite inspiring, talk about emergent four three-hour tutorials and hands-on
Mi Dai presented the Photometric LSST intelligence from neuromorphic complex- sessions that allowed the participants
Astronomical Time series Classification ity and synthetic synapses in nanowire to delve directly into the techniques. These
Challenge (PLAsTiCC) and described networks. After presenting a brief history covered an introduction to machine learn-
how to involve the community at large of AI, she showed that to reach the ing using Python notebooks, machine
and its thousands of machine learning ultimate goal of general intelligence, one learning and deep learning using disitrib-
experts via, for example, the Kaggle needs to move away from mainstream uted frameworks and optimised libraries,
platform 2, to come up with the best algo- computing. She told us that companies and how to use NIFTy.
rithms to classify the very many transient are already developing sophisticated
sources that will be found by the Large neuromorphic chips, which consume The workshop closed with a final discus-
Synoptic Survey Telescope. Similarly, orders of magnitude less power than sion led by Torsten Enßlin which proved
Rafael de Souza presented the Cosmos- conventional processors and which try to that AI is needed in astronomy and will be
tatistics Initiative , an endeavour aimed at emulate the brain. She also described even more in the future, especially as we
fostering interdisciplinary collaborations how scientists and engineers are now won’t be able to store all the data and
around astronomy and characterised by creating biomimetic structures of nano on-the-fly decisions will have to be made.
a residence programme. wires that self-assemble into a complex, It was also stressed that interdisciplinary
densely interconnected network, with a teams are required, as well as a new kind
It was also important to make sure that topology similar to a biological neural of physicist who will have to be trained.
the ground covered by the workshop was network and characterised by a collective The need to better understand the meth-
as wide as possible. Accordingly, John memory. ods that are used was also stressed —
Skilling spoke on how to do computation as scientists, we shouldn’t rely on “black
in big spaces, presenting the framework This very impressive series of invited talks boxes” and need to be very critical.
of inference, i.e. the Bayes therorem, and was complemented by numerous con- This requires us to learn the language of
how the prior space is often much bigger tributed talks and posters, covering the the data scientists and the basic under-
than the small posterior space, leading whole range of applications of AI meth- pinning of the methods, such as Bayesian
to a lot of confusion. Only by reducing ods in astronomy, from meteorite hunting probability.
dimensionality can one hope to solve the to augmenting N-body simulations with
problems. Jens Jasche showed how to deep learning models, through applica- The workshop was a great success and
perform large-scale Bayesian analyses tions in adaptive optics. A poster compe- participants praised the overall quality
of cosmological datasets, using computer tition was organised, and participants of the talks and tutorials, as well as of the
programs and not analytic functions to were asked to vote for the best posters. abstract booklet 4. Many were already
perform a hierarchical Bayes analysis. The three winners each received a hoping that a related workshop would
This allows one, for example, to infer the mounted ESO image; they were: Philipp take place next year! We therefore invite
mass density in a super-galactic plane or Baumeister (Using Mixture Density Net- the community to organise such events
estimate galactic cluster masses. In the works to Infer the Interior Structure as regularly as possible. The PDFs of all
same vein, Torsten Enßlin presented both of Exoplanets); Timothy Gebhard (Learn- the talks and posters and the material of
in a contributed talk and in a tutorial the ing Causal Pixel-Wise Noise Models two of the three tutorials are available on
fully Bayesian information field theory and to Search for Exoplanets in Direct Imag- the workshop webpage 1. All in all, the
the Numerical Information Field Theory ing Data); and Colin Jacobs (Using Deep talks, tutorials and posters covered a very
(NIFTy) library 3. Learning in the Cloud to find Strong wide range of topics in artificial intelli-
Lenses). gence and the workshop fulfilled its aim.
Perhaps even further from what we usu- The available material will surely be very
ally see in astronomical conferences, useful for many years to come.
Demographics we had 41% students, 22% postdoctoral shop, as well as ESO catering and ESO logistics,
for ensuring the best conditions during the meeting.
researchers, and 37% tenure-track
The workshop had a very high level of or tenured faculty. The talk selection was
participation, with about 130 registered made blindly (the chair of the SOC References
participants coming from all parts of the removed names and identifying informa-
Adorf, H.-M. 1991, The Messenger, 63, 69
world and approximately two dozen tion about the authors, including their
Baron, D. 2019, arXiv: 1904.07248
unregistered participants from ESO and seniority and their affiliation), and was Griffin, R. F. 2014, The Observatory, 134, 109
neighbouring institutes, including several based solely on the merits of the abstract Stoehr, F. 2019, ASPC, 387, 523
software engineers, highlighting the and its relation to the themes of the
great interest generated by the topic. workshop. This resulted in 62% of the
Notes
talks and 50% of the posters being given
The Scientific Organising Committee by students. a
achine learning is one of the most commonly
M
worked hard to ensure fair representa- used subsets of AI.
tion from the community. Among the
10 invited speakers, five were female. Acknowledgements
Links
Three of the five sessions were also It is a great pleasure to thank the members of the
chaired by women. Among the abstracts Scientific Organising Committee (Coryn Bailer-
1
orkshop website: https://www.eso.org/sci/
W
meetings/2019/AIA2019.html
submitted, a quarter were by women, Jones, Henri Boffin, Massimo Brescia, TorstenEnßlin, 2
T he Kaggle platform website: kaggle.com
and this was also the female/male ratio Emille Ishida, Zdenka Kuncic, Antoine Mérand, 3
Numerical Information Field Theory: http://ift.
Melissa Ness, and Felix Stoehr) for setting up an
among the contributed speakers. We amazing programme, the invited speakers for
pages.mpcdf.de/nifty/
had a very high level of participation from remarkable and clear reviews, and the organisers of
4
2019 AIA workshop programme: https://www.eso.
org/sci/meetings/2019/AIA2019/Booklet_final.pdf
young researchers, most likely due to a the four tutorials (Patrick van der Smagt, Fabio
combination of a highly discounted Baruffa, Luigi Iapichino, Philipp Arras, Torsten Enßlin,
Philipp Frank, Sebastian Hutschenreuter, and
registration fee for students and the fact Reimar Leike) for their exceptional work. We espe-
that this field is relatively young. Thus, cially thank Stella Chasiotis-Klingner for her efficient
among the registered participants, organisation of many practical aspects of the Work-
DOI: 10.18727/0722-6691/5181
Wolfgang Vieser 1 transfer from research institutes into the way to other scientific concepts, espe-
Tania Johnston 1 classroom is lacking. The goal of this cially in young people.
Saeed Salimpour 2, 3 conference was to bring together all
stakeholders — teachers, educators and Astronomy therefore plays a special role
researchers — to communicate and within public science communication.
1
ESO discuss their various needs in order to The literature is full of suggestions and
2
Deakin University, Burwood, Australia effectively bridge the gap between advice about how to best communicate
3
Edith Cowan University, Joondalup, astronomy education research and its astronomy to the public. Astronomy
Australia practical application. education or teaching astronomy is differ-
ent from communication, however.
Whereas communication and outreach
Astronomy education contributes to Astronomy is not only one of the oldest are processes aimed at generating inspi-
the spread of scientific literacy among sciences, but also a perennially fascinating ration and awareness, education aims
successive generations, helping to one to the broader public, who often ask to develop knowledge, skills and compe-
attract students into science, technology, educators questions such as, “where do tences, and core values and attitudes
engineering and mathematics (STEM) we come from?”, or “are we alone?” For through a range of pedagogies and
subjects and potentially also into astron- this reason, astronomy has always been methodologies that account for the abili-
omy research. Although the field of a relatively easy area of science to convey ties and development level of the learner.
research into astronomy education has to the public and it can serve as a gate- Astronomy education is less prominent
grown significantly, the sustainable within the scientific community than
astronomy communication and outreach Secondary Teacher Education — span how these could be organised into coher-
even though the International Astrono traditional and practical research, explor- ent patterns of understanding. A new
mical Union (IAU) established Commis- ing the purely theoretical issues encoun- AER study now provides a wider and
sion 46 on “The Teaching of Astronomy” tered when attempting to embed research more coherent framework about the high
in 1964. The Commission’s designation results into practical situations, usually conceptual understanding of astrophys-
changed from 46 to C1 in 2015 but its mediated by standards, curriculum and ics that is necessary to develop research-
mandate has remained essentially the instruction. based teaching-learning sequences for
same: to further the development and high school students — something that
improvement of astronomical education This conference was organised by IAU will be developed in the near future.
at all levels throughout the world through Commission C, together with ESO, the Other contributed talks focused on how
various projects developed and main- ESO Supernova, and Leiden University. It multidimensionality in the field of astron-
tained by the Commission and by dissemi- was hosted at the ESO Supernova using omy or astronomical time- and length
nating information concerning astronomy all its facilities, including the planetarium scales can be made understandable for
education. as the lecture theatre. The programme students. In both fields, models can help
comprised three invited talks, 44 contrib- students learn about relevant aspects,
To foster this mandate, the IAU will estab- uted talks, 10 hands-on workshops and but they need to be built by experienced
lish the Office of Astronomy for Education 50 posters. As it is an educational facility, teachers. Some contributed talks sur-
(OAE) this year; its objective will be to the ESO Supernova proved to be the veyed and analysed the production
provide structured support to for astron- perfect location for this conference and of AER studies in different countries like
omy education in all countries. This the participants were enthusiastic about Brazil, France, Japan and Portugal,
includes, but is not limited to, providing this inspiring environment. Details of the focusing not only on school grade levels
training and resources for encouraging programme can be found via the confer- or the type of academic research but also
the use of astronomy as a stimulus for ence webpage1. Each talk was followed on gender balance.
teaching and learning from primary to by a five-minute session dedicated to
secondary school levels. At a workshop questions and discussion that continued
between 17 and 19 December 2019 at during the breaks. Poster viewing took Astronomy education standards,
the Institut Astrophysique de Paris, the place during all coffee breaks and was curriculum and instruction
IAU revealed that the location of the OAE particularly encouraged during 30-minute
would be at the Haus der Astronomie in poster sessions every day. In his invited talk Robert Hollow dis-
Heidelberg, Germany. In addition, at that cussed opportunities and issues regard-
same workshop the remit of the new The IAU President Ewine van Dishoeck, ing curricula at the school level, particu-
office was presented along with its plans the IAU General Secretary Teresa Lago larly in the context of the recent IAU
regarding the goals set out in the IAU and ESO’s Director General Xavier Framework for Astronomy Literacy. The
Strategic Plan for 2020–2030. Barcons all acknowledged the necessity science curriculum of Australia served as
of such a conference in their welcome an example to illustrate the possibilities
The field of astronomy education has addresses. The IAU President also gave and the challenges of using astronomy in
grown significantly over the last few dec- a summary of the activities and events teaching science. A contributed talk by
ades, with an increasing number of commemorating 100 years of the IAU, Saeed S alimpour gave a review of how
research articles having been published including the travelling exhibition, and often astronomy is encountered in the
by a growing number of researchers. announced the inauguration of the OAE. school curriculum of 37 countries (OECD,
Despite this, there has been no regular China and South Africa), highlighting that
international conference for astronomy 77% of all curricula in Grade 1 include
education researchers and practitioners Astronomy education research astronomy, 54% in Grades 2 & 7 and
around the world to convene and discuss 27% in Grades 1 to 12. The highest per-
their work. This conference is intended The invited talk by Janelle Bailey sum centage of astronomy (85%) can be
to be the first of a regular, biennial, IAU- marised the broad field of astronomy found in Grade 6. The study also revealed
Commission C1 Astronomy Education education research (AER), highlight- that one curriculum explicitly mentioned
Conference series. The aim is to increase ing upcoming projects, for example a only two women astronomers and only
the quality, quantity, community and two-volume work about astronomy three of the 37 countries explicitly men-
impact of astronomy education research education, and introduced modern edu- tioned indigenous astronomy.
and practice by bringing together astron- cation concepts like active learning.
omers, astronomy education researchers Future directions of AER were also dis- Several contributed talks highlighted the
and education practitioners to communi- cussed, such as the use of qualitative importance of research-based science
cate, discuss and tackle common issues. and mixed methods, robust quantitative education, in which real data are ana-
analyses and longitudinal studies. lysed with research-quality tools to inves-
The three key themes of this conference tigate questions for which the answer is
— Astronomy Education Research; Contributed talks covered more special- not known. One talk recounted how the
Astronomy Education Standards, Curric- ised topics such as students’ (mis-)con- practices employed to use archival image
ulum and Instruction; and Primary and ceptions about astronomical topics and and spectral data have evolved over time
Internet resources like videos were shown The invited talk by Agueda Gras-Velazquez Another contributed talk described con-
to be extremely helpful for hearing- focused on the struggles of teachers in tinuous professional development work-
impaired or deaf people in a contributed their daily work; challenges include a lack shops for primary and secondary school
40%
Australia Malaysia 38%
Austria The Netherlands 43%
Belgium Poland 57%
Brazil Portugal
Canada Romania 62%
Chile Russia 60%
China Slovakia
France Spain
Germany Sweden
Hungary Thailand
Ireland United Kingdom
Italy USA
Japan
Figure 2. Pie chart showing the distribution of Figure 3. Multi-level pie chart showing the gender
c ountries from which the 114 participants came. ratio amongst participants (outer ring), talks (middle
ring), accepted posters (inner ring); in each of these
the lighter and darker colours represent the female
and male ratios, respectively.
teachers based around the Irish National – limitations and opportunities of plane- cated. After this meeting, the baseline
Junior Certificate theme of Earth and tariums, often seen as natural places to has now been set, and the community
Space. With these workshops, teachers run informal education activities; looks forward to marking its progress by
are kept informed of current research – how to make astronomy projects more the next conference in 2021.
and discoveries, and are provided with diverse and inclusive;
content and material to engage students – developing and testing new interdisci-
using space research. A further contrib- plinary and inclusive educational and Demographics
uted talk presented teacher trainings in outreach activities;
the use of robotic telescopes. The focus – links between astronomy and environ- The demand for this conference was
was to bring astronomy closer to teach- mental education; extremely high, but owing to the limited
ers in an enjoyable way, so that they lose – how to design inquiry-based work- seating in the planetarium, the number of
the fear of working on these topics with shops for secondary school students possible participants had to be capped
their students, and to provide them with and teacher trainings that are relevant at 114. The participants came from
the tools and knowledge so that they to curricula and cost effective. 25 countries, including 13 ESO member
can introduce them in a practical way and states, the Host Country Chile and
develop enquiry-based projects. The conference highlighted that astron- Strategic Partner, Australia (see Figures 1
omy education is a well-established field & 2). In total, 112 talk abstracts were sub-
with a global community. Education — mitted, 46% of which came from female
Workshops alongside research, outreach and devel- colleagues (Figure 3). The gender balance
opment — is one of the main activities among the speakers in the final pro-
The workshops covered a broad range of of the IAU. Some trends in astronomy gramme reflected the 40:60 (female:male)
activities: education recurred throughout the con- distribution of the participants, similar
– the positive and negative effects of the ference, such as multidisciplinary to that of the Scientific Organising Com-
use of technology in the classroom; approaches, the options for online collab- mittee (SOC), which had a corresponding
– the use of real astronomical data in the oration, training, and distribution. Also, ratio of 44:56 (female:male).
classroom; the societal relevance of education was
– presentation of the recently published addressed and discussed with topics
booklet Big ideas in Astronomy: A pro- like inclusion and diversity and climate Links
posed Definition of Astronomy Literacy; change. Very fruitful discussions took 1
C onference webpage: https://iau-dc-c1.org/
– t wo workshops combining STEM with place during the conference and the astroedu-conference/
the arts (STEAM), one dealing with pro- majority of participants felt that they are 2
AstroEdu platform: https://astroedu.iau.org/en/
grammable materials that are impor- acting towards a common goal. However,
tant for future space travel, the other there is still a need to improve knowledge
with the creative use of satellite images; transfer between researchers and practi-
– an art-based approach to teaching astron- tioners, as the wheel tends to be rein-
omy via Visual Thinking Strategies; vented too often, with efforts being dupli-
Fellows at ESO
notorious for clumsy accidents in the lab. Telescope on La Palma. During our five sis. Since the PN.S is a visitor instrument,
However, there was one topic I excelled nights at the telescope, we experienced we spent many afternoons leading up to
in and that was astrophysics. Unfortu- first-hand how it felt to be an astronomer our observations tuning the filters and
nately, the astronomy branch was closed and the patience it required in case of aligning the CCDs in the instrument arms.
in my second year of study. The subject bad weather! Yet I had found a new pas- Six months later, I got the opportunity
was not uppermost in my mind anymore, sion. It was rewarding to see our project to join my ESO Fellow mentor during his
and struggling with the prospect of grow from a little idea in our heads to duties at Paranal observatory. At last I
becoming a researcher, I seriously con- typing the coordinates of targets into the was convinced that the next step for me
sidered reverting to one of my earlier telescope, and to finally present the sci- would be an ESO Fellowship in Chile
career choices: becoming a teacher. I ence to our peers after reducing the data. to get even more exposure to the
had just made it to the state final of a One year later, I again found myself on forefront of astronomical research and
youth music competition in Germany and La Palma, this time observing at the instrumentation.
teaching music and physics in high William Herschel Telescope for my master
school seemed like the perfect combina- thesis project with Eline Tolstoy. And here I am now. I have just completed
tion of subjects for me. the first year of my fellowship and there-
It was clear that I wanted to pursue a fore the first 80 days and nights as a sup-
Everything changed, however, when I PhD in observational astronomy. In the port astronomer on Paranal. It has been
was selected for a summer internship at same year, I was accepted into the Inter- an exciting year with a steep learning
Mount Stromlo Observatory of the Aus- national Max Planck Research School curve! I am part of the Multi Unit Spectro-
tralian National University. For the first (IMPRS) on Astrophysics in Munich for a scopic Explorer (MUSE) instrument oper-
time, I got an insight into the day-to-day three-year studentship at ESO under the ations team and currently work on a
life of a researcher and could work inde- supervision of Magda Arnaboldi. For my project to investigate how well the adap-
pendently on a small project on stellar PhD, I investigated how the halos of tive optics improve the image quality. It
streams in the Milky Way. My supervisor early-t ype galaxies grow through mergers is great working in an international and
Ken Freeman introduced me to the beauty and accretion. This is a challenging interdisciplinary team. I particularly enjoy
and elegance of galaxy dynamics. All endeavour, as the closest early-type the ritual of watching the sunset from
of a sudden, I could appreciate classical galaxies are already millions of light-years the platform before the night starts. I also
mechanics as a great tool to describe the away, but the faint halos are very recently started to experiment with astro-
motions of the stars. After the internship, extended on the sky. I therefore use a photography. I like to share the wonders
I abandoned my idea to go to the con- particular type of stars — planetary of the night sky with my friends in the
servatory and instead focused on finding nebulae — which are like green beacons city, where due to the bright lights, one
an opportunity to carry out my bachelor in the sky, and whose velocity can be can barely make out the Southern Cross.
thesis research project in astronomy; so measured even at a distance of hundreds When I am not observing or working
I found a placement in nearby Groningen of millions of light-years. from Vitacura, one is likely to find me
to work with Amina Helmi. rehearsing music. While living in Munich,
I enjoyed being in the middle of one of I was a soprano with the Münchner
I decided to stay at the Kapteyn Institute the astronomy hubs in Europe and got to Motettenchor and spent a good part of
for another two years to complete my participate in many exciting seminars my leisure time in churches and concert
Master of Science, thoroughly enjoying a and conferences that were taking place halls in the region. Now in Santiago, I
curriculum centred on astronomy. Soon on campus. I travelled again to La Palma have again taken up singing, although on
an opportunity came up to enroll in a to observe the halos of giant elliptical a smaller scale. It is a relaxing balance
course on observational astronomy which galaxies with the custom-built Planetary to the academic world and a great way to
was to take place at the Isaac Newton Nebula Spectrograph (PN.S) for my the- practise my Spanish.
In Memoriam
ESO staff member, Cristian Herrera ator (TIO) in 2001. During his 18 years at 10 years, leading the night crew and
González, sadly passed away in August Paranal, Cristian worked on most of the was the coordinator of the Instrument
2019 and will be much missed. He joined telescopes, instruments and subsystems Operations Teams activities for the oper-
ESO and the Science Operations Depart- of the observatory. He held the role of ators during his shifts.
ment as Telescope and Instrument Oper- nighttime TIO Coordinator for more than
Personnel Movements
Europe Europe
Andersson Lundgren, Andreas (SE) Apex Support Astronomer Gentile Bordelon, Dominic (IE) Library Technology Specialist/
Fusillo, Nicola (IT) Fellow System Administration & Classification
Girdhar, Aishwarya (IN) Student IMPRS Specialist
Heida, Marianne (NL) Fellow Harrison, Christopher (UK) Fellow
Izquierdo Cartagena, Andrés (CO) Student DFG Heijmans, Jeroen (NL) Instrumentation Engineer/Physicist
König, Pierre-Cécil (FR) Student IMPRS Hellemeier, Joschua Andrea (DE) Student
Lansbury, George (UK) Fellow Hughes, Meghan (UK) Student
Marchetti, Tommaso (IT) Fellow Iani, Edoardo (IT) Student
Oliveira Teixeira, Emanuel Pedro (PT) Accountant Kolwa Sthabile, Namakau (ZA) Student IMPRS
Paredes, Amaya (ES) Technical Writer/ Lelli, Federico (IT) Fellow
Documentation Specialist Møller, Palle (DK) User Support Astronomer
Szakacs, Roland (AT) Student IMPRS Slater, Roswitha (DE) Administrative Employee
Teuber, Karin (DE) Administrative Assistant Zanella, Anita (IT) Fellow
Trovão Ferreira, Bárbara (PT) Public Information Officer
Chile Chile
Campana, Pedro (CL) Electronics Engineer Milli, Julien (FR) Operation Staff Astronomer
Dauvin, Louise (CL) System Engineer Silva, Karleyne (BR) Operation Staff Astronomer
De Rosa, Robert (UK) Operation Staff Astronomer Vogt, Frédéric (CH) Fellow
Farias, Cecilia (CL) Telescope Instruments Operator
Hsieh, Pei-Ying (TW) Fellow
Leon, Angelica (CL) Telescope Instruments Operator
Saint-Martory, Georges (FR) ELT Deputy Site Manager
Santamaría Miranda, Alejandro (ES) Fellow
Scicluna, Peter (UK) Fellow
Slumstrup, Ditte (DK) Fellow
DOI: 10.18727/0722-6691/5183
Erratum
P (selfreported | DeepThought)
2DMHNQHSX./"
0.75
DeepThought inferred knowledge
2DMHNQHSX4RDQR
Worst
0.0047 0.22 0.78
0.60
%Q@BSHNM
0.45
Median
Best
Expert General No knowledge
G #
XDS D@QR D@QR XD @
QR
/ @M X lX knowledge
- N RS G M
DS G
@ M
+D R VD D ,NQ Referee perceived knowledge
! DS
Figure 1. Scientific seniority distribution of the DPR Figure 5. Conditional probability for the various
sample (blue) compared to the ESO users sample combinations of self-reported and DeepThought-
(orange) from Patat et al. (2016). Note that the two inferred knowledge level.
central orange bars correspond to the middle
seniority class in Patat et al. (2016).
We would like to correct and update Fig- The Messenger, 177, 3). Figure 1 is the ure 5 has been updated and corrected.
ures 1 and 5 in The Distributed Peer same as previously published and only The rest of the article, its discussion and
Review Experiment by Patat et al. (2019, has an updated caption and labels, Fig- conclusions remain unchanged.