NIRPS Joins HARPS: Setting New Standards at Infrared Wavelengths

Artificial Intelligence Usage by ESO Telescope Users

The Promises and Challenges of the ALMA Wideband Sensitivity Upgrade
The Messenger
No. 194 | 2025
ESO, the European Southern Observatory, Contents
is the foremost intergovernmental astron-
omy organisation in Europe. It is sup- Astronomical Science
ported by 16 Member States: Austria, Fahrion, K., Lyubenova, M. – Exploring the Star Clusters in the Centres
­Belgium, the Czech Republic, Denmark, of Galaxies with MUSE 5
France, Finland, Germany, Ireland, Italy, Hammer, F., et al. – Young Stars Discovered in Dwarf Spheroidal Galaxies
the Netherlands, Poland, Portugal, Spain, Confirm their Recent Infall into the Milky Way 8
Sweden, Switzerland and the United
Kingdom, along with the host and partner Telescopes and Instrumentation
country of Chile and with Australia as a Doyon, R., et al. – NIRPS Joins HARPS: Setting New Standards
Strategic Partner. ESO’s programme is at Infrared Wavelengths 13
focused on the design, construction and Jones, I. J. – Optimisation of the SPHERE Adaptive Optics Setup at ~11 mag 19
operation of powerful ground-based Santos, N. C. – PoET: the Paranal solar ESPRESSO Telescope 21
observing f­acilities. ESO operates three Moehler, S., Freudling, W. – Flux Calibration for VLT and ELT Spectrographs 26
observatories in Chile: at La Silla, at
­Paranal, site of the Very Large Telescope, Astronomical News
and at Llano de Chajnantor. ESO is the Jerabkova, T., et al. – Distributed Peer Review at ESO:
European ­partner in the Atacama Large Demonstrating Success and Evolving Through Period 115 33
Millimeter/submillimeter Array (ALMA). Cur- Hilker, M. – ESO’s Scientific Visitor Programme in Garching 37
rently ESO is engaged in the construction Nazari, I., et al. – Usage of Artificial Intelligence by ESO Telescope Users 39
of the Extremely Large ­Telescope. De Breuck, C., Díaz Trigo, M. – Report on the ESO workshop
“The Promises and Challenges of the ALMA Wideband Sensitivity Upgrade” 41
The Messenger is published in electronic Jerabkova, T., et al. – Report on the ESO workshop “Galaxies at Crossroads” 44
form twice per year. ESO produces and Miotello, A., Macias, E. – Report on the ESO workshop
distributes a wide variety of media con- “New Heights In Planet Formation” 47
nected to its activities. For further infor- Emsellem, E., et al. – Report on the ESO workshop
mation, contact the ESO Department of “A decade of Discoveries with MUSE and Beyond” 49
Communication at: De Simone, M., Drevon, J., Artur de la Villarmois, E. – Fellows at ESO 53

ESO Headquarters
Karl-Schwarzschild-Straße 2
85748 Garching bei München, Germany
Phone +498932006-0
[email protected]

The Messenger
Editor: Mariya Lyubenova
Editorial assistant: Isolde Kreutle
Copy-editing and proofreading:
Peter Grimley
Graphics, Layout, Typesetting:
Lorenzo Benassi
Online Publishing: Mafalda Martins
Design: Jutta Boxheimer
messenger.eso.org Front Cover: This image, taken with the VLT Survey
Telescope hosted at ESO’s Paranal Observatory,
Unless otherwise indicated, all images in shows the beautiful nebula NGC 6164/6165. The
The Messenger are courtesy of ESO, nebula is a cloud of gas and dust surrounding a pair
of stars called HD 148937. In a new study using ESO
except authored contributions which are data, astronomers have shown that the two stars are
courtesy of the respective authors. unusually different from each other — one appears
much younger and, unlike the other, is magnetic.
© ESO 2025 Moreover, the nebula is significantly younger than
either star at its heart, and is made up of gases nor-
ISSN 0722-6691 mally found deep within a star and not on the outside.
These clues together helped solve the mystery of the
The Messenger and all articles are pub- HD 148937 system — there were most likely three
lished open access under a Creative Com- stars in the system until two of them clashed and
merged, creating a new, larger and magnetic star.
mons Attribution 4.0 International License. This violent event also created the spectacular nebula
that now surrounds the remaining stars.
Credit: ESO/VPHAS+ team. Acknowledgement: CASU

2 The Messenger 194 | 2025

 S. Otarola/ESO

A solitary antenna points timidly at the Moon. But

this is not some lonesome telescope, but one of
the 66 antennas that together make up the impres-
sive Atacama Large Millimeter/submillimeter Array
(ALMA), operated by ESO and its international
partners.

The Messenger 194 | 2025 3

Astronomical Science

ESO/VVVX survey

This is an 80-million-pixel picture of the star

­cluster RCW 38, located 5500 light-years away in
the constellation Vela. RCW 38 is a young cluster
containing about 2000 stars, and is bursting with
star-forming activity.
Astronomical Science DOI: 10.18727/0722-6691/5377

Exploring the Star Clusters in the Centres of Galaxies

with MUSE

Katja Fahrion 1 Multi Unit Spectroscopic Explorer (MUSE) MUSE observations of FCC47 (NGC 1336),
Mariya Lyubenova 2 instrument at the VLT has allowed us to a nucleated elliptical galaxy in the Fornax
circumvent these limitations. With MUSE cluster at a distance of 20 Mpc (Figure 1).
data of nearby (< 50 Mpc) galaxies it has Line-of-sight velocities of 24 GCs and
1
Department of Astrophysics, University become possible to extract and analyse metallicities of five GCs were measured.
of Vienna, Austria the spectra of star clusters nestled within Fahrion et al. (2020b) then applied this
2
ESO the central regions of galaxies. As the method to data of 32 Fornax galaxies
host galaxy can be studied from the that were observed as part of the ESO
same data, a direct comparison between Large Programme Fornax3D (Sarzi et al.,
Massive star clusters are ubiquitous the stellar population and the kinematic 2018). In total, 733 GCs with reliable
in the central regions of galaxies. For properties of the host galaxy and its star velocity measurements were found. For
example, nuclear star clusters are pres- clusters becomes possible. With such an a subsample of 238 GCs metallicity
ent in most galaxies, and bulge regions approach, the inner star cluster systems measurements were also possible. With
can host globular clusters. Even though of galaxies can be explored, and even the this sample, a non-linear translation
these star clusters are bright, studying formation pathways of nuclear star clus- between Hubble Space Telescope colours
their properties is limited by the under- ters can be unveiled. and metallicities was found (Fahrion et
lying galaxy light. Here we discuss how al., 2020c) and spectroscopic catalogues
integral-field spectroscopy with the Multi at larger radii from multi-object spectros-
Unit Spectroscopic Explorer (MUSE) Probing inner globular cluster systems copy could be added (for example,
has enabled studies of the inner globular Chaturvedi et al., 2022). Moreover, with
cluster systems of massive galaxies GCs are dense star clusters, characterised this sample it was possible to test how
and how MUSE has allowed us to con- by old stellar ages, which makes them well GCs trace the properties of the
strain the formation mechanisms of powerful tracers of galaxy assembly and underlying host galaxy. Comparing the
nuclear star clusters. evolution. GCs within the bulge region rotation amplitude and velocity dispersion
of the Milky Way have recently been dis- of the GC systems with the rotation and
cussed as fossil remnants of bulge forma- dispersion of the host galaxies, it was
Introduction tion (Ferraro et al., 2021), but less is shown that the red GCs in particular are
known about the GC populations in the good tracers of the motion of galaxy
The central regions of galaxies are home inner regions of massive galaxies. spheroids. Additionally, comparing GC
to many morphological structures, such as metallicities with the host’s metallicities
discs, bars or bulges, shaped by various Fahrion et al. (2019) described the at the projected positions of the GCs
processes. Within these structures, mas- approach of extracting and analysing star showed that these red GCs also follow
sive star clusters such as nuclear star cluster spectra from MUSE data using the metallicity profile of the host.
clusters (NSCs) and globular clusters
(GCs) can be embedded. Even though FCC 47
NSCs and GCs are both dense star clus-
ters with millions of stars packed tightly 8
together — and therefore inherently
bright — studying those objects within
the central regions of galaxies is challeng- 7
ing, owing to the underlying galaxy back- Nuclear star cluster
ground. In photometric studies, the galaxy 6
light is often modelled and subtracted to
derive the colours and sizes of NSCs and
GCs; however, such an approach is not 5
y (kpc)

possible with slit or multi-object spectros-

copy. For this reason, spectroscopic stud- 4
ies of GC systems are usually limited to
the outer regions of galaxies (for example,
3 Globular clusters
Forbes et al., 2017), and slit spectroscopy
of NSCs is preferably done on bulgeless
spirals or faint dwarf galaxies, where it 2
is assumed that the host galaxy is not
contributing significantly to the light in the
1
centre (for example, Paudel, Lisker &
Kuntschner, 2011; Kacharov et al., 2018).
Figure 1. MUSE image
0 of FCC 47 in the Fornax
The high spatial sampling combined with 0 2 4 6 8
cluster with NSC and
the wide field of view provided by the x (kpc) GCs highlighted.

The Messenger 194 | 2025 5

Astronomical Science Fahrion, K., Lyubenova, M., Exploring Star Clusters in Centres of Galaxies with MUSE

Constraining nuclear star cluster created when NSCs form through the centre, suggesting a formation from GCs
formation rapid in-spiral of star clusters formed spiralling into the centre. Building on this,
in the central region that then spiral in Fahrion et al. (2021) then presented a
The formation of NSCs can be a complex directly (for example, Guillard, Emsellem larger sample of 25 early-type galaxies,
process and typically two main pathways & Renaud, 2016), which might be an mainly in Fornax, spanning a range of
are discussed (see Neumayer, Seth & important channel at high redshift. galaxy masses from 107 to 1011 M⊙.
Böker, 2020 and references therein, and Considering the metallicity differences
Figure 2): (i) formation through star forma- On the other hand, formation through the between NSCs and hosts as well as NSC
tion directly in the galaxy centre, following mergers of GCs is a singular way to explain star formation histories, they found a
the accretion and compression of gas, and metal-poor populations within NSCs, clear transition in the dominant NSC for-
(ii) via the mergers of massive star clusters. for example as in the Milky Way (Do et al., mation (Figure 3). In low-mass galaxies
While these star clusters can be young 2020). This channel, in its purest form, (< 10 9 M⊙), NSCs were found to be old
clusters formed very close to the galaxy only considers the dry merger of GCs and metal-poor and were likely formed
centre, traditionally the inspiral of GCs has and therefore no additional star formation through GC in-spiral, while in massive
been considered. In this way, NSC forma- is considered. As such, the NSC formed galaxies in-situ formation can explain
tion might be connected to GCs. is expected to reflect the properties of their high masses, complex star forma-
GCs, which are characterised by old popu- tion histories and high metallicities. Inter-
The in-situ or central star formation sce- lations and low metallicities. estingly, indications of both formation
nario depends on the mechanisms to channels were found for intermediate-­
funnel gas into the central region. Forma- mass galaxies.
tion directly in the galaxy centre then has Dominant NSC formation channel
consequences for the NSCs formed in Regardless of this clear result, with only
that way. For example, it can explain the While there are indications that the NSC galaxies in a galaxy cluster the question
presence of very young stars seen in formation channel depends on galaxy arose whether this trend of NSC forma-
the NSC of the Milky Way (for example, mass and type (see Neumayer, Seth & tion from GCs in dwarf galaxies would
Schödel et al., 2020), and the formation Böker, 2020 for a discussion), to under- hold up in star-forming dwarfs. To address
from gas through dissipative processes stand this process in individual galaxies this, Fahrion et al. (2022) presented
explains the sometimes elongated, rotat- the properties of NSCs and their hosts a novel sample of nine late-type dwarfs
ing and young NSCs in nearby spiral gal- must be compared. observed with MUSE. Even in this sample,
axies (Seth et al., 2006). As this process the NSCs were found to be mainly old
depends on the star formation activity Using a similar approach as for the GCs, and metal-poor, and the contribution
in the galaxy centre, the NSC formed we can use MUSE data to study the stel- from additional in-situ star formation was
can show a complex, extended star for- lar population properties of NSCs from small. This further confirmed that the
mation history and can reach metallicities background-cleaned spectra and com- NSCs in dwarf galaxies form from GCs
exceeding those of typical GCs as a pare them to the properties of the under- and therefore closely resemble GCs in
result of being formed from already pre- lying host galaxy. This allows us to com- their properties. However, the star forma-
enriched gas and metal-retention in the pare, for example, the NSC metallicity tion history of the galaxy imprints addi-
deep potential well of the galaxy centre. with that of the host. As a first example tional populations onto the NSC, which
Complex star formation histories and that this approach can unveil the domi- makes NSCs important records of past
high metallicities, however, can also be nant NSC formation mechanism, Fahrion star formation episodes.
et al. (2020a) studied two dwarf spheroi-
Figure 2. The two most discussed NSC formation dal galaxies observed with MUSE. In
channels: formation through the accretion and inspi- both cases, very metal-poor NSCs were Conclusions
ral of GCs (left), and formation directly in the galaxy found, even less enriched than the host
centre through in-situ star formation (right). In the
galaxies. In the case of one dwarf galaxy, The MUSE instrument at the VLT has
latter, young star clusters might be formed first
in the central region and then quickly spiral in (see KK 197, the NSC even shares its low changed how spectroscopic studies of
the zoom-in). metallicity with a GC found near the star clusters can be conducted. This is
seen beyond the works mentioned in this
article, as similar approaches have been
Globular cluster accretion Central star formation used, for example, to study planetary
GC nebulae in the central regions of galaxies
(for example, Spriggs et al., 2021) or to
Infall of young clusters explore globular clusters in dwarf galax-
ies (for example, Müller et al., 2020).
Moreover, recent work has employed
methods similar to those described here
to analyse a larger sample of nucleated
galaxies (Lyu et al., 2024), further confirm-
ing the trend with galaxy mass.

6 The Messenger 194 | 2025

 Nevertheless, unanswered questions
NSC formation channel
remain. For example, it is unclear when 8.5
Central star formation
NSCs and the observed trends in their for-
8.0 Both
mation pathways are established. Are the
GC accretion
NSCs we see today already formed from 7.5
then proto-GCs at high redshift? Or do the

log(MNSC/Mც)
GCs form first and then merge in the later 7.0
evolution? To answer these questions,
6.5 Figure 3. Dominant
the fraction of galaxies with a nuclear star NSC formation channel
cluster at higher redshift would be as function of galaxy
6.0
needed, but no such observations have and NSC mass. Dwarf
yet been made. Additionally, it is unclear galaxies form their
5.5
NSCs predominantly
why some galaxies have GCs but no NSC. through the accretion of
The lack of NSCs in high-mass galaxies 5.0 GCs, while the massive
might be explained by interactions with NSCs in massive galax-
supermassive black holes (SMBHs), but 7 8 9 10 11 ies form most of their
mass through central
the dynamical friction timescales are so log(Mgal/Mც)
star formation.
short in dwarfs that NSCs formed through
GCs should be ubiquitous. Perhaps gal- ones, involving many subsequent mergers Fahrion, K. et al. 2020c, A&A, 637, A27
Fahrion, K. et al. 2021, A&A, 650, A137
axy interactions or the underlying dark of galaxies and their NSCs and/or SMBHs.
Fahrion, K. et al. 2022, A&A, 667, A101
matter profile might hinder such an in-­ Ferraro, F. R. et al. 2021, Nat. Astron., 5, 311
spiral, but conclusive results even for Forbes, D. A. et al. 2017, AJ, 153, 114
individual systems are still missing (for Acknowledgements Guillard, N., Emsellem, E. & Renaud, F. 2016,
MNRAS, 461, 3620
example, Meadows et al., 2020). Another
KF acknowledges funding from the European Kacharov, N. et al. 2018, MNRAS, 480, 1973
avenue is to couple detailed orbit-based Union’s Horizon Europe research and innovation Lyu, W. et al. 2024, arXiv:2412.03132
dynamical models of galaxy nuclei with programme under the Marie Skłodowska-Curie Meadows, N. et al. 2020, MNRAS, 491, 3336
stellar population parameters. Looking into grant agreement No. 101103830. Müller, O. et al. 2020, A&A, 640, A106
Neumayer, N., Seth, A. & Böker, T. 2020,
the orbital distribution of NSCs can give us
A&A Rev., 28, 4
important hints about the evolutionary References Paudel, S., Lisker, T. & Kuntschner, H. 2011, MNRAS,
history of the galaxy nucleus. In turn, this 413, 1764
can enlighten us about whether the scaling Chaturvedi, A. et al. 2022, A&A, 657, A93 Sarzi, M. et al. 2018, A&A, 616, A121
Do, T. et al. 2020, ApJL, 901, L28 Schödel, R. et al. 2020, A&A, 641, A102
relations between galaxy nuclei and host
Fahrion, K. et al. 2019, A&A, 628, A92 Seth, A. C. 2006, AJ, 132, 2539
galaxy properties are driven by physical Fahrion, K. et al. 2020a, A&A, 634, A53 Spriggs, T. W. et al. 2021, A&A, 653, A167
processes, like AGN feedback, or statistical Fahrion, K. et al. 2020b, A&A, 637, A26
ESO/Iodice et al.

This image shows a pair

of overlapping spiral gal-
axies, NGC 3314a and
b, in the top left, caught
in a majestic cosmic
dance — captured by
ESO’s VLT Survey
Telescope (VST).

The Messenger 194 | 2025 7

Astronomical Science DOI: 10.18727/0722-6691/5378

Young Stars Discovered in Dwarf Spheroidal Galaxies

Confirm their Recent Infall into the Milky Way

François Hammer 1 galaxies reached the MW halo less than Gaia observations (the second and third
Piercarlo Bonifacio 1 3 Gyr years ago. This completely changes data releases, DR2 & DR3) provided
Elisabetta Caffau 1 the interpretation of their dynamics, detailed orbital motions for 156 globular
Yanbin Yang 1 mass and dark matter content. clusters (GCs; Vasiliev, 2019), and for
Frédéric Arenou 1 46 MW dwarf galaxies (Li et al., 2021),
Carine Babusiaux 2 allowing their orbital (or binding) energies
Monique Spite 1 A recent infall of most dwarf galaxies is to be accurately calculated. Several stud-
Patrick François 1 predicted by the hierarchical scenario ies (Kruijssen et al., 2019, 2020; Massari,
Ana Gomez 1 Koppelman & Helmi, 2019; Malhan et al.,
David Katz 1 In the last 30 years, astronomers have 2022) showed that several GCs are asso-
Lorenzo Monaco 3 conducted extensive observations and ciated with the important accretion events
Marcel Pawlowski 4 analyses of the stellar populations in the that occurred in the MW, namely the
Jianling Wang 1 dwarf galaxies surrounding the Milky Way elaboration of the bulge (12–13 Gyr ago),
(MW) galaxy. Several dwarf galaxies were the Kraken (11–12 Gyr ago) and the
thought to be made up of only very old
1
LIRA, Paris Observatory – PSL, CNRS, stars (with ages much greater than 6 Gyr)
France with a low concentration of elements Figure 1. The left panel shows total energy versus
2
Grenoble Observatory, CNRS, France heavier than helium. It was then deduced angular momentum (h = RGC x V tan) on a logarithmic
3
Andres Bello University, Concepciòn, that these dwarf galaxies, such as the scale, for high-surface-brightness GC (crosses),
low-surface-brightness GC (filled circles), and dwarf
Chile Sculptor dwarf spheroidal (dSph), had
4 galaxies (triangles). Structures identified by Malhan
Leibniz Institute for Astrophysics, AIP, lost their gas at these remote epochs, et al. (2022) and Kruijssen et al. (2020) are added in
Potsdam, Germany when they became satellites of our different colours. VPOS (Vast POlar Structure; see
Galaxy, orbiting around it ever since. Pawlowski, Pflamm-Altenburg & Kroupa, 2012 and
Li et al., 2021) dwarf galaxies are shown in blue.
This scenario has a major consequence
The dot-dashed line shows the limit that cannot be
Recent observations from ESA’s Gaia in near-field cosmology: these dwarf passed by any orbits, as it is fixed for a circular orbit
satellite and with ESO’s Very Large ­galaxies must have a huge quantity of (the largest possible binding energy for a given
Telescope have identified the presence dark matter in order to protect their stellar angular momentum). The right panel shows the
­c orresponding lookback time of stellar system entry
of a population of young stars, 0.5 to content from the disruptive force of the
in the Milky Way halo as a function of its current
2 Gyr old, in the halo of, and in dwarf MW’s gravitational field. Indeed, in the binding energy for different families of GCs, and for
spheroidal galaxies surrounding, the absence of a large amount of dark matter, the dwarf galaxies that do not belong to the tightly
Milky Way (MW). It suggests that MW tidal forces from the MW would disperse bound Sgr system (that is, excluding Sgr, Segue I, II,
Tucana III, IV and Willman I). The blue solid line is a
dwarf galaxies, currently devoid of gas, the stars of the dwarf galaxy in just a few
linear fit. A simple interpretation is that dwarf satel-
had, until recent times, enough gas to Gyr. Until now, they have been considered lites with log(Ebinding/(km2 s –2)) < 4.34 are on their
sustain a burst of star formation. The as the most dark-matter-­dominated ­initial approach (see blue box), a value close to the
recent loss of gas coincides with their ­galaxies in the Universe, whose total logarithm of the average energy (4.14 km2 s –2) of
dwarf galaxies, whose scatter provides an upper
arrival in the vicinity of the MW, in masses, derived from their large velocity
limit of Ebinding = 4.34. The latter combined with the
agreement with orbital predictions from dispersions, are 10 to 1000 times larger linear fit suggests a lookback time of halo entry of
Gaia that indicate that most dwarf than their stellar masses. dwarf galaxies less than 3 Gyr.

5.5
Bulge
First elaboration of the bulge
Bulge (12–13 Gyr)
Kraken
Kraken
Kraken (11–12 Gyr) GSE
5 GSE merger(s)
log(E binding ) (km 2 s –2)

Sgr
GSE (8–10 Gyr)
Sgr infall
Sgr (4–6 Gyr)
4.5

Dwarf galaxies
4
Dwarf

2 3 4 Lookback time (Gyr)

log (h) (km s –1 kpc)

8 The Messenger 194 | 2025

Figure 2. Left: UVES spectrum of the young metal-­ 1.2
poor star GHS143, for which the two Fe I lines indi-
cate a low-metallicity star. The blue line is the syn- Scl12 (Fe/H) = –0.32
thetic spectrum corresponding to the average
metallicity and the red line is the best fit to each line.
The high proper motions imply that the star is
unbound and falling into the Galaxy, its estimated (Fe/H) = –1.8
age being between 5 and 9 Myr. Right: GIRAFFE
spectra of two young stars in Sculptor, for which the
868.86 nm Fe I line indicates a difference of over 1.0
1 dex in iron abundance. The spectrum of Scl12 has
been arbitrarily shifted by –0.4, for display purposes.
The blue line is the synthetic spectrum correspond-
ing to the average metallicity, while the red line is
the best fit to the line.

Gaia-Sausage-Enceladus (GSE, 8–10 Gyr 0.8

ago) merger events, and more recently
R.I.

the dwarf Sagittarius galaxy (Sgr, 4–6 Gyr

ago) infall into the Galactic halo. These GHS143 Scl2 (Fe/H) = –1.54
associations have been determined after
comparing GC ages, metallicities and the
orientation of their orbital angular momenta
and binding energy to those of stars from 0.6
which the above events have been identi-
fied. They are expected since stars and
GCs are often formed together through
gas-rich major-merger events (De Lucia
et al., 2024; Valenzuela et al., 2024).

Figure 1 shows the distribution of GCs 0.4

633.5 633.7 868 869 870
(crosses for the high-surface-brightness
Wavelength (nm) Wavelength (nm)
examples, large dots for the low-surface-­
brightness ones; see Hammer et al., 2021)
and of dwarf galaxies (triangles) in the most of the latter galaxies have reached blue straggler, because of its anomalous
binding energy–angular momentum plane. the MW halo less than 3 Gyr ago. position in the colour–magnitude diagram.
GCs are coloured on the basis of the However, the formation of blue stragglers
structures to which they belong. Each with masses larger than two solar masses
structure (bulge, Kraken, GSE, and Sgr) Discovery of young stars in the halo is unlikely because most halo stars are
shows a very narrow range in binding and then in dSph galaxies > 8 Gyr old, for which the maximum mass
energy, which allows them to be identified at the turn-off is expected to be 0.9 M⊙.
in the relation between binding energy Thanks to the proper motions provided To explain young and massive halo stars
and infall lookback time. by Gaia, Bonifacio et al. (2024) and through the blue straggler channel would
­Caffau et al. (2024a,b) identified several require mergers involving three stars,
Galaxies like the MW follow a hierarchical stars in the MW halo with high velocities which are extremely rare events. The sim-
structure formation, in which smaller with respect to the Sun (> 500 km s –1). plest explanation of the data is that these
­galaxies merge into larger systems over 10 of these high-velocity stars show large stars are not rejuvenated, but are genu-
time. This ‘inside-out’ growth pattern masses (in excess of 1.3 M⊙, half of them inely young.
means that older structures are tightly in excess of 2 M⊙), young ages (0.3 to
bound to the galaxy, while more recent 2.5 Gyr old) and metallicities from –1.3 The origin of these halo young stars is a
arrivals are less so. Hammer et al. (2023) to –2.2. Figure 2 shows the Ultraviolet puzzle, since the halo does not contain
derived the MW’s accretion history (see and Visual Echelle Spectrograph (UVES) large amounts of gas to fuel star forma-
Figure 1), which can be fitted by a single spectrum of GHS 143 with a mass of tion. It is thus likely that they are formed
line from bulge elaboration to Sgr. The 2.3 M⊙. Measurements of metallicity from during the earlier infall of a dwarf galaxy,
line slope is in agreement with predictions spectroscopy combined with accurate for which the original gas content has
from high-resolution cosmological simu- Gaia photometry allow an accurate deter- been ram-pressurised during its interac-
lations (Rocha, Peter & Bullock, 2012; mination of the star’s mass and age a. It is tion with the Galactic corona. It has
and see a detailed comparison to simu­ well known that some stars may ‘rejuve- prompted us to verify whether or not
lations in Hammer et al., 2024a). By nate’ by mass transfer from a binary com- there might be also a population of young
extrapolating the line to the lower binding panion or even by merging with another stars in the most massive dwarfs, i.e.,
energy of dwarf galaxies, it suggests that star; such a rejuvenated star is called a dSph galaxies. Yang et al. (2024) have

The Messenger 194 | 2025 9

Astronomical Science Hammer, F., et al., The Young Stellar Content of dSph Galaxies around the Milky Way

T = 0.00 Gyr T = 0.75 Gyr T = 1.30 Gyr T = 2.10 Gyr

Dwarf motion Dwarf gas (cyan dots) is compressed by the Galactic corona inducing star
to the Galaxy formation and supernovae, which explosions expell the gas to the dwarf outskirts

T = 3.25 Gyr T = 3.60 Gyr T = 4.55 Gyr T = 4.85 Gyr

Optical observation

Galactic corona density increases during the fall of the dwarf, and it becomes very
effective at removing progressively the dwarf gas and freeing the stars (orange dots)
Fading to observation depth

been able to ‘filter’ the dSph members by fairly common, especially because other Figure 3. Snapshots of a video summarising the
hydrodynamical simulation of the infall of a gas-rich
excluding the foreground stars (from the dSphs (Fornax, Carina, Leo I and Leo II)
dwarf galaxy that is ram-pressurised during its infall
MW) that appear in the field of view of the are also known to contain young popula- into the MW corona (Wang et al., 2024). Gas is repre-
dSph galaxies. This filtering is extremely tions. The most favourable location in sented by cyan particles and stars by red particles.
efficient, and was not possible before the the colour-magnitude diagram to identify The yellow arrow indicates the dwarf motion, and the
insert on the top-left the radial evolution with time.
availability of the Gaia data. For example, young stars is the so-called ‘yellow
The final galaxy in the bottom-right panel has proper-
in the field of view of Sculptor, half of the plume’ (see, for example, Gullieuszik et ties very similar to the Sculptor dSph.
stars are foreground, but thanks to the al., 2008), an almost vertical sequence to
filtering using the Gaia proper motions, the blue of the red giant branch. These
this contamination is at most 1.4%! This stars are three to four magnitudes brighter stars are known to form within gas clouds,
filtering method allowed the unambiguous than the TO and are young evolved stars the dwarf galaxies must have contained
identification of young stars in Sculptor in the core helium-burning phase, or sizable amounts of gas up to 0.5 to 2 Gyr
whose ages are between 0.5 and 2 Gyr young sub-giant stars. The same region years ago. It is well accepted that MW
and whose masses are up to three times can also be occupied by evolved blue dwarf progenitors were gas-rich galaxies
that of the Sun, after determining their stragglers, but such an interpretation similar to present-day dwarf irregular
metallicities through observations with becomes unlikely for stars of mass above ­galaxies observed in the field. Once they
the GIRAFFE spectrograph (see Figure 2). two solar masses. arrived in the MW halo, their gas was
Besides confirming that these stars stripped by the ram pressure exerted by
belong to Sculptor, spectroscopic obser- the hot gas in the MW corona. The pro-
vations allow their masses to be derived. Conclusions cess is quite rapid, because masses of
Because the turnoff (TO) point of the dwarf galaxies are three to five orders of
Sculptor dSph is similar to that of the halo The discovery of young stars in four dSph magnitude smaller than that of our galaxy.
(0.9 M⊙), it is also unlikely that Sculptor galaxies formerly considered as uniquely
young stars can be blue stragglers. made of old stars changes our view of The above discoveries have considerable
Besides Sculptor, three additional dSphs, the history of their motions relative to the consequences for our understanding of
Sextans, Ursa Minor and Draco, show MW. It supports the scenario of a recent the nature of dwarfs surrounding the MW.
the presence of a young star population, infall suggested by the precise determina- Hydrodynamical simulations (Wang et al.,
demonstrating that the phenomenon is tion of their orbits from Gaia. In fact, since 2024) show that during the interaction

10 The Messenger 194 | 2025

 with the MW corona gas, the stellar content in Sculptor and other dwarf galaxies are Chiti, A. et al. 2023, AJ, 165, 55
De Lucia, G. et al. 2024, MNRAS, 530, 2760
is considerably shaken by turbulence needed. Young stars with masses in
Gullieuszik, M. et al. 2008, MNRAS, 388, 1185
effects during the process (see Figure 3 b). excess of two solar masses have rota- Hammer, F. et al. 2021, ApJ, 922, 93
dSph progenitors gradually lost their gas, tional velocities in excess of 200 km s –1 Hammer, F. et al. 2023, MNRAS, 519, 5059
provoking a strong disequilibrium in the when on the main sequence. As they Hammer, F. et al. 2024a, A&A, 692, L1
Hammer, F. et al. 2024b, MNRAS, 527, 2718
residual stellar content, including by tidal evolve they slow down, but they can still
Kruijssen, J. M. D. et al. 2019, MNRAS, 486, 3180
shocks exerted by the MW’s gravity show measurable rotational velocities of Kruijssen, J. M. D. et al. 2020, MNRAS, 498, 2472
(Hammer et al., 2024b). When the gas is the order of 10–20 km s –1 (Lombardo Li, H. et al. 2021, ApJ, 916, 8
fully lost, stars begin to expand thanks to et al. 2021). Evolved blue stragglers, Lombardo, L. et al. 2021, A&A, 656, A155
Longeard, N. et al. 2022, MNRAS, 516, 2348
the associated loss of internal gravity, and instead, can spin up to 100 km s –1 at the
Malhan, K. et al. 2022, ApJ, 926, 107
this naturally explains the presence of time of mass transfer, but spin down Massari, D., Koppelman, H. H. & Helmi, A. 2019,
stars associated with many dwarfs while ­rapidly, even more so as they evolve. At A&A, 630, L4
being very far from their centres (Chiti et the same time we need to obtain high-­ Pawlowski, M. S., Pflamm-Altenburg, J. & Kroupa, P.
2012, MNRAS, 423, 1109
al., 2023; Sestito et al., 2023; Longeard resolution, high-signal-to-noise spectra
Rocha, M., Peter, A. H. G. & Bullock, J. 2012,
et al., 2022; Waller et al., 2023). Hydrody- for the many halo yellow-plume stars, that MNRAS, 425, 231
namical simulations predict that, in the we can select from Gaia. Their metallicity Sestito, F. et al. 2023, MNRAS, 525, 2875
case of a recent arrival, a small number and dynamics should trace their origin Valenzuela, L. M. et al. 2024, A&A, 687, A104
Vasiliev, E. 2019, MNRAS, 484, 2832
of stars must be formed during the inter- to an existing or disrupted dwarf galaxy.
Wang, J. et al. 2024, MNRAS, 527, 7144
action between the gas-rich progenitor Waller, F. et al. 2023, MNRAS, 519, 1349
and the MW corona, which is consistent If confirmed, this novel paradigm for Yang, Y. et al. 2024, A&A, 691, A363
with the observed number of young stars explaining the observations of MW dwarf
(Yang et al., 2024). They also reproduce galaxies will become a serious contender
Notes
all the properties of dwarf galaxies, includ- to the scenario of dark matter-dominated
ing their observed velocity dispersions, MW dwarf galaxies, and this could lead a
 etermination of metallicity through spectroscopy
D
with a very limited amount of dark matter, us into a new area in near-field cosmology. allows the appropriate isochrones in the colour–
magnitude diagram to be chosen, and hence the
or even without dark matter at all.
stars’ ages and masses to be estimated.
b
A video of the simulation of the Sculptor dSph
Our results are suggestive and require References
(Wang et al., 2024) shown in Figure 3 can be seen at
further confirmation. In particular spectro- Bonifacio, P. et al. 2024, A&A, 684, A91
https://www.youtube.com/watch?v=SwxSdmfQis4
scopic observations at higher signal-to- Caffau, E. et al. 2024a, A&A, 683, A72
noise ratio of more yellow plume stars Caffau, E. et al. 2024b, A&A, 684, L4
B. Haeussler/ESO

In this portrait of ESO’s Paranal Observatory, taken through them. This is due to their orbital paths being the line that connects them crosses the Milky Way
in early February, our planets appear to parade relatively aligned along a single, flat plane called the at an angle. This is due to the ecliptic being tilted at
one after the other across the night sky. In addition ecliptic. (In reality, the planets aren’t aligned one after about 60 degrees to the galactic plane on which our
to the Moon, our own Milky Way, and the comet the other in a straight line in the Solar System, they entire Milky Way lies. If the Milky Way could some-
C/2024 G3, we can see Saturn, Venus, Jupiter and are fanned out; but we can still see them simultane- how be shrunk down to lie flat on a table, our Solar
Mars — even Neptune and Uranus are hiding here too! ously in the sky, which only happens every few years.) System would be jutting out like a pin stuck in it at
Often on nights with a few planets in view, you can You may notice that in this image the planets are not an odd angle.
draw an imaginary straight line in the night sky contained within the band of the Milky Way, and that

The Messenger 194 | 2025 11

Telescopes and Instrumentation

A. de Burgos Sierra/ESO

This impressive image of comet C/2024 G3 (ATLAS)

was captured on 29 January from ESO’s Paranal
Observatory by Abel de Burgos Sierra, an ESO
Fellow in Chile. Gas and dust particles are ejected
from the nucleus and pushed away from the Sun
by the solar wind and radiation, creating a spectac-
ular display with multiple tails.

12 The Messenger 194 | 2025

Telescopes and Instrumentation DOI: 10.18727/0722-6691/5379

NIRPS Joins HARPS: Setting New Standards at

Infrared Wavelengths

René Doyon 1 Ryan Cloutier 21 Mirsad Sarajlic 17

François Bouchy 2 João Coelho 15, 16 Alex Segovia 2
Francesco Pepe 2 Marion Cointepas 2, 4 Julia Seidel 3, 37, 2
Claudio Melo 3 Uriel Conod 2 Damien Ségransan 2
Étienne Artigau 1 Nicolas B. Cowan 22, 23 Robin Schnell 2
Lison Malo 1 Eduardo Cristo 9, 10 Ana Rita Costa Silva 9, 10, 2
François Wildi 2 João Gomes da Silva 9 Avidaan Srivastava 1, 2
Frédérique Baron 1 Laurie Dauplaise 1 Atanas K. Stefanov 6, 7
Xavier Delfosse 4 Antoine Darveau-Bernier 1 Marcio A. Teixeira 5
Jose Renan De Medeiros 5 Roseane de Lima Gomes 1, 5 Stéphane Udry 2
Jonay I. González Hernández 6, 7 Daniel Brito de Freitas 24 Dania Valencia 19
Nuno C. Santos 9, 10 Elisa Delgado-Mena 25, 9 Valentina Vaulato 2
Gregg Wade 11 Jean-Baptiste Delisle 2 Joost P. Wardenier 1
Romain Allart 1 David Ehrenreich 2, 26 Bachar Wehbe 15, 16
Khaled Al Moulla 2 João Faria 2, 9 Drew Weisserman 21
Nicolas Blind 2 Pedro Figueira 2, 9 Ivan Wevers 14
Charles Cadieux 1 Dasaev O. Fontinele 5 Vincent Yariv 4
Bruno L. Canto Martins 5 Thierry Forveille 4 Gérard Zins 3
Neil J. Cook 1 Jonathan Gagné 27, 1
Xavier Dumusque 2 Ludovic Genolet 2
Yolanda Frensch 2, 3 Félic Gracia Témich 6 1
University of Montreal, Mont-Mégantic
Frédéric Genest 1 Olivier Hernandez 27 Observatory and Trottier Institute
Nolan Grieves 2 Melissa J. Hobson 2 for Research on Exoplanets, Canada
Gaspare Lo Curto 3 Jens Hoeijmakers 28, 2 2
Geneva Observatory, Department
Christophe Lovis 2 Norbert Hubin 3 of Astronomy, University of Geneva,
Lucille Mignon 2, 4 Farbod Jahandar 1 Switzerland
Louise D. Nielsen 2, 3, 12 Ray Jayawardhana 29 3
ESO
Anne-Sophie Poulin-Girard 13 Hans-Ulrich Käufl 3 4
Grenoble Alpes University, CNRS,
José Luis Rasilla 6 Dan Kerley 14 IPAG, France
Rafael Rebolo 6, 7, 8 Johann Kolb 3 5
Department of Theoretical and
Vladimir Reshetov 14 Vigneshwaran Krishnamurthy 22 Experimental Physics, Federal University
Danuta Sosnowska 2 David Lafrenière 1 of Rio Grande do Norte, Brazil
Micheal Sordet 2 Pierrot Lamontagne 1 6
Canary Islands Institute of Astrophysics
Jonathan Saint-Antoine 1 Pierre Larue 4 (IAC), Spain
Alejandro Suárez Mascareño 6, 7 Henry Leath 2 7
Department of Astrophysics, University
Simon Thibault 13 Alexandrine L’Heureux 1 of La Laguna, Spain
Philippe Vallée 1 Izan de Castro Leão 5 8
Spanish National Research Council
Thomas Vandal 1 Olivia Lim 1 (CSIC), Spain
Manuel Abreu 15, 16 Allan M. Martins 5, 2 9
Institute of Astrophysics and Space
José L. A. Aguiar 5 Jaymie Matthews 30 Science, University of Porto, CAUP,
Guillaume Allain 13 Jean-Sébastien Mayer 1 Portugal
Tomy Arial 1 Yuri S. Messias 1, 5 10
Department of Physics and Astronomy,
Hughes Auger 13 Stan Metchev 31 Faculty of Science, University of Porto,
Susana C. C. Barros 9, 10 Leslie Moranta 1, 27 Portugal
Luc Bazinet 1 Christoph Mordasini 17 11
Department of Physics and Space
Björn Benneke 1 Dany Mounzer 2 Science, Royal Military College
Xavier Bonfils 4 Nicola Nari 32, 6, 7 of Canada, Kingston, Canada
Anne Boucher 1 Ares Osborn 21 12
University Observatory, Faculty
Vincent Bourrier 2 Mathieu Ouellet 1 of Physics, Ludwig-Maximilians
Sébastien Bovay 2 Jon Otegi 2 University, Munich, Germany
Christopher Broeg 17, 18 Léna Parc 2 13
Centre of Optics, Photonics and
Denis Brousseau 13 Luca Pasquini 3 Lasers, Laval University, Canada
Vincent Bruniquel 2 Vera M. Passegger 6, 7, 33, 34 14
Herzberg Astronomy and Astrophysics
Martha Bryan 19 Stefan Pelletier 2, 1 Research Centre, National Research
Alexandre Cabral 15, 16 Céline Peroux 3 Council of Canada
Andres Carmona 4 Caroline Piaulet-Ghorayeb 1, 35 15
Institute of Astrophysics and Space
Yann Carteret 2 Mykhaylo Plotnykov 19 Science, Faculty of Science, University
Zalpha Challita 1, 20 Emanuela Pompei 3 of Lisbon, Portugal
Bruno Chazelas 2 Jason Rowe 36

The Messenger 194 | 2025 13

Telescopes and Instrumentation Doyon, R., et al., NIRPS Joins HARPS

16 30
Department of Physics, Faculty of University of British Columbia, Canada an overview of the design of NIRPS, its
31
Science, University of Lisbon, Portugal Western University, Department of on-sky performance, its Guaranteed
17
Space Research and Planetary Physics & Astronomy and Institute for Time Observation programme, and its
Sciences, Physics Institute, University Earth and Space Exploration, Canada first scientific results.
32
of Bern, Switzerland Light Bridges S.L., Teide Observatory,
18
Center for Space and Habitability, Spain
33
University of Bern, Switzerland Hamburg Observatory, Germany Introduction
19 34
Department of Physics, University Subaru Telescope, National Astronomical
of Toronto, Toronto, Canada Observatory of Japan, USA The discovery of the first exoplanet orbit-
20 35
Aix Marseille University, CNRS, CNES, Department of Astronomy & Astrophysics, ing a solar-type star (Mayor and Queloz,
LAM, Marseille, France University of Chicago, USA 1995), and of the first transiting exoplanet
21 36
Department of Physics & Astronomy, Department of Physics & Astronomy, (Charbonneau et al., 2000), stand as
McMaster University, Canada Bishop’s University Canada ­pivotal moments in astrophysics. The quest
22 37
Department of Physics, McGill Lagrange Laboratory, Côte d’Azur for nearby habitable worlds and evidence
University, Canada Observatory, CNRS, Côte d’Azur of biological activity beyond the Solar
23
Department of Earth & Planetary University, France System has prompted the construction
Sciences, McGill University, Canada of powerful observatories such as Kepler
24
Department of Physics, Federal (Koch et al., 2010), the Transiting
University of Ceará, Brazil The Near-InfraRed Planet Searcher Exoplanet Survey Satellite (TESS; Ricker
25
Centre for Astrobiology (CAB), (NIRPS) is a high-resolution, near-­ et al., 2015), the James Webb Space
CSIC-INTA, ESAC campus, Spain infrared spectrograph optimised for ­Telescope (JWST; Gardner et al., 2023)
26
Life in the Universe Center, Faculty detecting and characterising exoplanets and, very soon, ESO’s Extremely Large
of Science, University of Geneva, around low-mass stars, working in Telescope (ELT; de Zeeuw, Tamai &
Switzerland ­tandem with the High Accuracy Radial Liske, 2014).
27
Montreal Planetarium, Espace pour velocity Planet Searcher (HARPS). While
la Vie, Canada HARPS set new standards 20 years ago
28
Lund Observatory, Division of with its metre-per-second-level precision,
Figure 1. Left: schematic of the NIRPS frontend.
Astrophysics, Department of Physics, NIRPS follows this successful path, Right: images from the guiding camera of a binary star
Lund University, Sweden achieving even better precision at infra- (0.3 arcseconds separation) with the AO loop open
29
York University, Canada red wavelengths. This article presents (top) and closed (bottom). From Bouchy et al. (2025).

0.4ೀ sky/ref
Frontend 0.4ೀ science tet01 Mic I = 4.73
0.4ೀ sky/ref tet02 Mic J = 5.84
n

Motorised 0.9ೀ science

h e ecisio

cy

stage
he ficen
ad
r
hP

ad

Fibre injection optics

f
Hig

hE
Hig

M3 fold
nm

nIR guiding camera

70 0–24 00

FP3
Pierced
mirror
Spectral calibration (2×)

To calib system

Motorised
NIRPS AO loop open
Bonded stage (2×)
Harps/ collimator
Separation 0.3 arcsec
NIRPS
ADC
dichro
m
0n
HARPS 360–690 nm

0
24

WFS / science dichro WFS camera

FP2
0–
98

700–970 nm

DM241

3.6 m Cass NIRPS AO loop open

focus
Differential Tip-Tilt Mount (motorised 2 DOF) AO system

14 The Messenger 194 | 2025

 Figure 2. Extracted and wavelength-calibrated
50000 ­s pectrum of Proxima Cen in HE mode, centered at
1268 nm. Blue: uncorrected spectrum; orange:
telluric-­corrected spectrum. From Bouchy et al. (2025).
40000
Extracted flux (e –)

the precision of RV measurements and

30000 significantly improves our ability to dis-
entangle planetary signals from stellar
­jitter noise. This article presents a brief
20000 overview of NIRPS, including its design,
on-sky performance, initial results, and
10000 Telluric uncorrected highlights from its extensive Guaranteed
Telluric corrected Time Observation (GTO) programme.
1264 1266 1268 1270 1272 For a more comprehensive description
of NIRPS, readers are referred to Bouchy
et al. (2025).
The rapid progress in exoplanet research advantages have driven the development
over recent decades was driven largely of high-resolution NIR spectrographs,
by precision velocimetry and in particular including GIANO (Oliva et al., 2012), NIRPS+HARPS: a unique dual optical-­
the development of fibre-fed optical spec- the Habitable-zone Planet Finder (HPF; infrared precision velocimeter
trographs like the High Accuracy Radial Mahadevan et al., 2012), the Calar Alto
velocity Planet Searcher (HARPS; Mayor high-Resolution search for M dwarfs with NIRPS is a fibre-fed, highly-stabilised
et al., 2003), which set new standards by Exoearths with Near-infrared and optical echelle spectrograph operating in the
achieving metre-per-second precision. Échelle Spectrographs (CARMENES; NIR and installed on the ESO 3.6-metre
Advances in large-format infrared detec- Quirrenbach et al., 2014), the InfraRed telescope at La Silla, Chile. The instru-
tors, largely motivated by and developed Doppler (IRD) instrument (Kotani et al., ment includes a frontend bonnette at the
for JWST, paved the way for a new 2018), and SpectroPolarimètre InfraRouge Cassegrain focus, linked via optical fibres
­generation of precision infrared spectro- (SPIRou; Donati et al., 2020). to the cryogenic spectrograph in the
graphs tailored to studying low-mass stars. coudé room. The frontend integrates an
This context prompted the Universities The Near-InfraRed Planet Searcher adaptive optics (AO) system to enhance
of Montreal and Geneva, in collaboration (NIRPS), the newest addition to this suite, efficiency and minimise the instrument’s
with the Institute of Astrophysics and builds on the legacy of HARPS (Mayor et size. As for HARPS, NIRPS’s spectrograph
Space Science (Portugal), the Canaries al., 2003) and the Echelle SPectrograph is housed in a thermally-controlled enclo-
Institute of Astrophysics (Spain), Grenoble for Rocky Exoplanet and Stable Spectro- sure to ensure optimal thermal stability.
Alpes University (France), the Federal scopic Observations (ESPRESSO; Pepe
­University of Rio Grande do Norte (Brazil), et al., 2021), while leveraging the infrared
and ESO, to initiate the development expertise and experience gained from
of an ‘infrared HARPS’ for the southern SPIRou. Operating in tandem with HARPS,
Figure 3. High cadence of NIRPS RV detrended data
hemisphere. NIRPS provides an unparalleled optical-­ of Proxima Cen, along with the best model fit, show-
NIR capability for precision velocimetry. ing residuals with an RMS of 81 cm s –1. From Suárez
Low-mass M dwarfs, which dominate the This dual-­wavelength approach enhances Mascareño et al. (2025).
Milky Way’s stellar population (Reylé et al.,
2021), are excellent targets for exoplanet
studies. Their small radii and masses 4
amplify detection signals via radial velocity
(RV) and transit methods, while their low
luminosity means that habitable zones 2
lie closer to the star, with orbital periods
measured in weeks rather than a year,
∆ RVDet (ms –1)

0
greatly simplifying the characterisation of
potentially habitable exoplanets.
–2
The near-infrared (NIR) is particularly
suited to M dwarf studies, as it mitigates –4
stellar activity jitter on RV measurements
compared to the optical and provides
access to helium and molecular signa- –6
tures like H2O, O2, CO, CH4, and CO2, 10340 10360 10380 10400 10420 10440
critical for atmospheric studies. These BJD – 2450000 (d)

The Messenger 194 | 2025 15

Telescopes and Instrumentation Doyon, R., et al., NIRPS Joins HARPS

NIRPS operates in the Y, J, and H bands,

covering a wavelength range from 2023-06-23
972.4 nm to 1919.6 nm. It offers two fibre 4
2023-07-28
sizes (0.4 and 0.9 arcseconds) yielding 2023-08-24
resolving powers of R = 90 000 for the Average
high-accuracy (HA) mode and R = 75 000 3
for the high-efficiency (HE) mode. This
spectral range enables the detection of Excess absorption (%)
molecular signatures such as water and 2
methane in planetary atmospheres. The
frontend module (Figure 1) houses the
AO system, which corrects for atmospheric 1
turbulence to improve light coupling into
the fibres under variable seeing condi-
tions. A dichroic beam splitter simultane-
0
ously directs light to both HARPS and
NIRPS, enabling parallel optical and NIR
observations.
–1
The fibre link transports light to the
spectrograph and incorporates a fibre 10828 10830 10832 10834 10836
stretcher and double scrambler to reduce Wavelength (Å)
modal noise that impacts radial velocity
(RV) measurements. The calibration unit
includes hollow-cathode (HC) uranium-­ 1.02
neon lamps and a Fabry-Pérot (FP) étalon
to illuminate a reference fibre alongside
1.01
the science fibre. The HC provides abso-
lute wavelength calibration, while the FP
allows drift correction. Calibrations are 1.00
performed daily, with a laser frequency
Normalised flux

comb currently under commissioning. 0.99

The backend spectrograph is housed

0.98
within a vacuum vessel, with its optical
bench stabilized to 75 K, maintaining
thermal variations within 0.1 mK. The 0.97
optical design features a reflective dou-
ble-pass collimator, a 13-lines mm –1 R4 0.96
echelle grating, a carousel of five ZnSe
prisms for cross-dispersion, and a refrac- 0.95
tive camera that feeds a 4096 × 4096-
pixel Hawaii-4RG (H4RG) infrared detec-
tor with 15-μm pixels. The full wavelength –0.02 –0.01 0.00 0.01 0.02 0.03
range spans 71 orders, with the line Phase
spread function sampled by three pixels.
Bouchy et al., 2025). The two independent Figure 4. Transmission spectra (top) and excess
helium light curves (bottom) comparison between
pipelines are particularly valuable for cross-­
transits (in cyan, orange, and rose) and the average
Data pipeline validation and assessing the robustness in black. On the top panel, three spectral regions
of scientific results. (blue, green, and red) are identified to measure the
The NIRPS data reduction pipeline temporal variation of the helium signature. From
Allart et al. (2025).
(NIRPS-DRS) is adapted from the The pipeline begins with order localisation,
ESPRESSO pipeline, incorporating fea- flat-fielding, and wavelength calibration.
tures inspired by the APERO pipeline The FP etalon ensures precise drift moni-
(Cook et al., 2022), a versatile framework toring when necessary a, while uranium-­ runs. A major challenge in NIR precision
initially developed for SPIRou. While neon lamps, combined with FP frames, velocimetry is the contamination of stellar
NIRPS-DRS and APERO share similarities, provide an estimate of the FP cavity spectra by telluric absorption lines. The
they differ in key aspects, such as their length, forming the basis for the absolute NIRPS-DRS pipeline includes a telluric
approaches to telluric correction (see calibration across nights and observing subtraction module developed for

16 The Messenger 194 | 2025

ESPRESSO (Allart et al., 2022), while RV performance was characterised on frequently planet formation occurs around
APERO is using an ensemble of telluric several RV standards with known planetary such stars with masses below 0.1 M☉;
standards, fast-rotating early-­t ype stars, systems such as Proxima Cen, featuring and (3) understanding the process of
and a principal component analysis two planets including an Earth-mass one planet formation and dynamical evolution
(PCA)-based method for modelling and in the habitable zone (Proxima b). As shown by searching for planets around young,
removing telluric lines. Figure 2 displays in Figure 3, Proxima b is clearly detected very low-mass stars.
a portion of the J-band spectrum of with a residual noise of ~80 cm s –1 com-
­Proxima Cen, both before and after telluric pared to 2.5 m s –1 from the HARPS data
correction, demonstrating the effective- alone. NIRPS is the first NIR velocimeter to Mass characterisation of transiting
ness of our method, which is absolutely demonstrate sub-­metre-per-second perfor- planets orbiting M dwarfs
crucial for achieving metre-per-second- mance, partly due to the excellent sub-­
level precision at infrared wavelengths. Kelvin thermal stability of the spectrograph This programme is dedicated to providing
yielding typical drifts of 3–4 cm s –1 day–1 mass measurements of transiting
Radial velocity extraction is performed and wavelength uncertainties at the level exoplanets unveiled by TESS and other
using two complementary methods: the of 50–70 cm s –1. transit surveys through various sub-
standard cross-correlation function (CCF) programmes. Mass measurements are
technique and the line-by-line (LBL) High-resolution spectroscopy is a power- essential for interpreting transmission
method (Artigau et al., 2022). The CCF ful tool for probing exoplanetary atmos- spectra obtained with JWST and con-
provides immediate RV measurements for pheres via transmission spectroscopy straining internal structure models. This
each observation, while the LBL method, and constraining orbital architecture, programme aims to shed light on the
which constructs a template spectrum including the spin-orbit angle, through the nature, formation and evolution of super-
from a time series, generally delivers Rossiter–McLaughlin (RM) effect. This Earths and mini-Neptunes. One sub-­
more precise RVs with less sensitivity to capability was demonstrated by observ- programme is dedicated to precise (~10%)
outliers, achieving, for example, a twofold ing three transit events of the warm mass measurements of small rocky planets
improvement on the ESPRESSO uncer- Saturn WASP-69b. As shown in Figure 4, to constrain their core mass fraction.
tainties for the temperate super-Earth NIRPS successfully detected the helium
LHS 1140b (Cadieux et al., 2024). triplet near 1083 nm in the planet’s
atmosphere, with evidence of variability High-resolution spectroscopy of
indicative of cometary-like tail mass loss. exoplanet atmospheres
On-sky performance and first results The RM measurements suggest a slightly
misaligned orbit. The third core programme centres on
NIRPS commissioning took place over atmospheric studies of exoplanets, pri-
two years from November 2019 to March marily hot gas giants, using transmission
2023 with the official first light on 17 May NIRPS Guaranteed Time Observation and emission spectroscopy. Its objective
2022. The commissioning phase of programme highlights is to uncover the chemistry, dynamics,
NIRPS demonstrated excellent perfor- and orbital architectures of exoplanet
mance, meeting or exceeding its design In exchange for building and operating atmospheres. The programme combines
requirements. the instrument, ESO awarded the NIRPS a broad atmospheric reconnaissance
consortium 725 nights over five years. survey with in-depth analyses of a care-
The instrument’s overall throughput peaks These are allocated to three core science fully selected sample of exoplanets, aim-
at 13% in the H band. The AO system programmes, each receiving 225 nights, ing to establish critical reference datasets
significantly improves fibre coupling effi- with an additional 50 nights reserved for in preparation for the upcoming ELT era.
ciency, achieving typical encircled energy ‘other science’ programmes.
of 55% and 70% for the HA and HE modes,
respectively; this performance is constant Other science
up to I = 11. Modal noise is mitigated Blind RV search for exoplanets orbiting
through the fibre stretcher and through AO nearby low-mass stars HARPS and NIRPS provide a unique
scanning by using the tip/tilt mirror to move capability, each performing at the metre-
the star within the fibre core randomly This core sub-programme is primarily per-second level, for stellar activity studies
during an exposure. Together, both stretch- dedicated to a blind search for planets as well as stellar characterisation, includ-
ing and AO-scanning yield a modal noise around M dwarfs (< 0.6 M⊙) with three ing abundance determination (for exam-
reduction by a factor of five for the HE major objectives: (1) identifying the near- ple, Jahandar et al., 2025), in particular
mode, leaving a residual noise of 0.43% est exoplanetary systems amenable to refractory elements (Fe, Mg, Si) which are
(SNR~230) similar to the flat-field stability atmospheric characterisation in reflected critical inputs for internal structure model-
of 0.65% which translates into RV noise light with the ELT (Snellen et al., 2015; ling. This approach was applied to LHS
of 0.9 m s –1. The AO was successfully Pallé et al., 2023) and which are orbiting 1140 using commissioning data, revealing
tested to lock onto small (< 2-arcsecond) M dwarfs within approximately 6 pc; that the temperate super-Earth LHS 1140b
Solar System objects, such as Saturn’s (2) searching for exoplanets around is likely a water world with a 10–20% water
and Jupiter’s moons. nearby ultra-cool dwarfs to estimate how mass fraction (Cadieux et al., 2024).

The Messenger 194 | 2025 17

 Telescopes and Instrumentation Doyon, R., et al., NIRPS Joins HARPS

Solar observations With its exceptional performance and Cadieux, C. et al. 2024, ApJL, 960, L3
Charbonneau, D. et al. 2000, ApJL, 529, L45
ambitious GTO programme, NIRPS is
Cook, N. J. et al. 2022, PASP, 134, 114509
The HARPS Experiment for Light Integrated poised to play a central role in exoplanet de Zeeuw, T., Tamai, R. & Liske, J. 2014,
Over the Sun (HELIOS) solar telescope research. NIRPS lays the foundation and The Messenger, 158, 3
(Dumusque et al., 2015) feeds both expertise needed for the ELT era, where Donati, J.-F. et al. 2020, MNRAS, 498, 5684
Dumusque, X. et al. 2015, ApJL, 814, L21
HARPS and NIRPS, and continuously high-resolution infrared spectroscopy will
Gardner, J. P. et al. 2023, PASP, 135, 068001
monitors the Sun as a star. High-cadence be essential for characterising the atmos- Jahandar, F. et al. 2025, ApJ, 978, 154
solar spectra enable detailed insight into pheres of nearby exoplanets through Koch, D. G. et al. 2010, ApJL, 713, L79
solar variability and its effect on disc-­ reflected light. Kotani, T. et al. 2018, Proc. SPIE, 10702, 1070211
Mahadevan, S. et al. 2012, Proc. SPIE, 8446,
integrated radial velocity and on the
84461S
retrieval of planetary atmospheric param- Mayor, M. & Queloz, D. 1995, Nature, 378, 355
eters (Mercier et al., 2025). Acknowledgements Mayor, M. et al. 2003, The Messenger, 114, 20
Mercier, S. et al. 2025, submitted to A&A
The NIRPS project became a reality thanks to the Oliva, E. et al. 2012, Proc. SPIE, 8446, 84463T
support and vision of ESO, the scientific leadership Quirrenbach, A. et al. 2014, Proc. SPIE, 9147, 91471F
Summary of NIRPS’s cohesive consortium, and key financial Paille, E. et al. 2023, submitted to Exp Astron,
contributions from the Canada Foundation for arXiv:2311.17075
Innovation, the Trottier Family Foundation, the Swiss
NIRPS represents a major milestone in National Science Foundation, the Spanish Ministry
Pepe, F. et al. 2021, A&A, 645, A96
Reylé, C. et al. 2021, A&A, 650, A201
precision infrared velocimetry, achieving of Science, the French National Research Agency, Ricker, G. R. et al. 2015, JATIS, 1, 014003
sub-metre-per-second precision through the Fundação para a Ciência e a Tecnologia (FCT) in Snellen, I. et al. 2015, A&A, 576, A59
advanced telluric subtraction techniques Portugal and the Brazilian funding agency CNPq. Suárez Mascareño, A. et al. 2025, submitted to A&A
and post-processing algorithms like LBL.
Early results demonstrate its considerable References Notes
potential for precise mass determination,
exoplanet atmospheric characterisation, Allart, R. et al. 2022, A&A, 666, A196 a
In practice, NIRPS is so stable that the FP is not
Allart, R. et al. 2025, accepted to A&A
and stellar studies, including activity Artigau, É. et al. 2022, AJ, 164, 84
used during science observations.
­analysis and abundance measurements. Bouchy, F. et al. 2025, submitted to A&A
ESO/Digitized Sky Survey 2. Acknowledgment: Davide De Martin

This image shows the

RCW 38 star cluster
in visible light. Dust
absorbs most light at
these wavelengths,
hiding large areas of
this cluster from us.

18 The Messenger 194 | 2025

Telescopes and Instrumentation DOI: 10.18727/0722-6691/5380

Optimisation of the SPHERE Adaptive Optics Setup

at ~11 mag

Matías I. Jones 1

1
ESO

We have extended the recent analysis

of the SPHERE adaptive optics (AO)
performance for faint stars. In particular,
we compared the raw contrast reached
using the medium-frequency (600 Hz)
and low-frequency (300 Hz) modes on
different targets with G > 11 mag, and
under different atmospheric conditions.
We found that using the medium-­
frequency mode in this magnitude range 1.0
f = 300 Hz; τ0 = 4.3 ms; seeing = 0.73ೀ
leads to significantly better contrast.
f = 600 Hz; τ0 = 4.2 ms; seeing = 0.82ೀ
Based on these results, we have updated
Normalised raw contrast

0.8
the AO frequency setup accordingly,
that is, we extended the 600 Hz mode
by one magnitude, up to 11.5 mag. 0.6

0.4
Introduction

The Spectro-Polarimetric High-contrast 0.2

Exoplanet REsearch instrument (SPHERE;
Beuzit et al., 2019) is a high-contrast 0.0
imaging facility that is equipped with a 200 300 400 500 600 700 800 900 1000
powerful adaptive optics (AO) system, Angular separation (mas)
called SAXO (Fusco et al., 2006). The
instrument comprises three different sub- Figure 1. IRDIS coronagraphic raw images of V* NN
Lup observed on 2023-06-16, using the 300 Hz and
systems: the Zurich imaging polarimeter
the 600 Hz modes (upper left and right image,
(ZIMPOL), which is sensitive to visible respectively). The corresponding normalised raw
light; the InfraRed Dual-band Imager and contrast curves are shown below.
Spectrograph (IRDIS) and the integral
field spectrograph (IFS), that are sensis-
tive to near-IR light (Y, J, H and K bands); by Jones et al. (2022; hereinafter J22). tion, pushing the SAXO limits. Motivated
and SAXO, which is equipped with a This limits our ability to study stars in by this, and based on the results presented
­visible wavefront sensor (WFS). SPHERE young stellar associations, which are in J22, we tested different setups to opti-
has proven to be efficient at detecting intrinsically red and hence significantly mise the AO correction and to maximise
faint companions and protoplanetary discs, fainter in the visible than in the near-IR. the scientific return of SPHERE data.
reaching a final contrast better than 15 mag This problem will be addressed by the
(beyond ~ 300 mas), when using angular SPHERE AO system upgrade, called
and reference differential imaging (for SPHERE+ (Bocaletti et al., 2020), which Observations
example, Wahhaj et al., 2021) on bright will be equipped with a pyramid WFS
targets, and reaching a Strehl ratio greater sensitive to IR light. Hence, before this We observed three stars on four different
than 90%. However, one of the main upcoming upgrade (whose first light is nights, under different atmospheric con-
­limitations of SAXO is the use of a visible foreseen for 2027), it is very important ditions a. Details of the observations are
WFS, whose performance is strongly to optimise the AO setup, so that we are summarised in Table 1. We mainly followed
degraded beyond G ~ 12 mag, as shown able to reach the best possible AO correc- the methodology presented in J22, that

300 Hz 600 Hz
Star name G mag Night τ0 (ms) seeing (arcsec) τ0 (ms) seeing (arcsec)
V* NN Lup 11.6 2022-08-01 4.8 0.8 4.8 0.9
V* NN Lup 11.6 2023-06-16 4.4 0.5 4.5 0.5
Table 1. List of the observed stars, Gaia EDR3 G-band UCAC2 18885095 11.3 2024-04-20 4.5 0.6 5.5 0.7
magnitude, night of observations, and mean atmos-
pheric conditions during the observing sequence. 2MASS J13015435-4249422 11.1 2022-0731 4.5 0.8 4.7 0.8

The Messenger 194 | 2025 19

 Telescopes and Instrumentation Jones, I. J., Optimisation of the SPHERE Adaptive Optics Setup at ~11 mag

Figure 2. H-band raw contrast as a function of the

UCAC2 18885095 instant flux received in the visible WFS. The dots and
V* NN Lup (2023-06-16) crosses correspond to individual frames observed
V* NN Lup (2022-08-01) with the 300 Hz and 600 Hz modes, respectively.
2MASS J13015435-4249422 The dashed line corresponds to the measured con-
trast for the 300 Hz mode, including observations
10 –2 taken under TCAT30 (τ0 > 4.5 ms; seeing < 0.80
H-band 5σ raw contrast

arcsec) or better atmospheric conditions (updated

from Figure 3 in J22).

10 –3

10 –4
1 2 3 4 5 6 7 8 9 10 11 12 13
Flux WFS (e –/subaperture/frame)

is, the observations were performed with can be seen in the raw images. In par- a relatively conservative choice, consider-
IRDIS, using the H23 filter. For each star ticular, the effect of the wind-driven halo ing that we can expect a similar contrast
we first obtained a flux (F) sequence (see Cantalloube et al., 2020) is partially for stars as faint as G ~ 12 mag, where
­followed by an object (O) coronagraphic supressed with the 600 Hz mode. More- the degradation of the AO correction is
sequence. The detector integration time over, the contrast curves show a signifi- expected to take place (corresponding to
for each flux frame was between the cant improvement with this mode, up to a about 3 e – per subaperture per frame).
­minimum of 0.8 s and 8 s, and for the separation of about 500 mas. Finally, Fig- We plan to re-adapt this limit accordingly.
object frames was 32 s. We repeated the ure 2 shows the raw contrast at 300 mas, Finally, these tests show that for future
procedure using the same star, but this as a function of the instant flux measured instruments equipped with extreme AO
time changing between the low-frequency by the visible WFS, for the two different systems, such as SPHERE+ itself, a more
(f = 300 Hz) and the medium-­frequency modes. As can be seen, there is a signifi- flexible AO frequency correction should
(f = 600 Hz) AO modes. We did this cant improvement (typically a factor of be selected based on the instant flux on
immediately after the first F–O sequence around two to three) in the raw contrast the WFS.
was finished, so that the airmass and achieved when using the 600 Hz mode.
atmospheric conditions were as close
as possible between the two sets of Acknowledgements
observations. Summary and conclusions MJ acknowledges the help of Eduardo Peña to per-
form these tests, and Zahed Wahhaj for discussion
Motivated by the results reported in J22, of the results.
Methods and results we have compared the SPHERE/IRDIS
H-band raw contrast achieved when using References
For each individual frame in each the 300 Hz and 600 Hz AO frequency
sequence, we computed the normalised modes, for stars with G ~ 11 mag. For Beuzit, J.-L. et al. 2019, A&A, 631, A155
raw contrast curve, and the 5-σ raw con- this, we performed back-to-back obser- Bocaletti, A. et al. 2020, arXiv:2003.05714
Cantalloube, F. et al. 2020, A&A, 638, A98
trast at a separation of 300 mas (see vations of stars in this magnitude regime, Fusco, T. et al. 2006, Proc. SPIE, 6272, 62720K
equation 1 of J22). To visualise the differ- swapping between these two modes. Jones, M. I. et al. 2022, A&A, 667, A114 (J22)
ence in the AO correction using these As expected, we observed a significant Wahhaj, Z. et al. 2021, A&A, 648, A26
two modes, we compared individual improvement in the raw contrast when
frames and their resulting contrast curves. using the 600 Hz mode, in all four observ- Notes
Figure 1 shows an example of this, for ing sequences. Based on these results,
the G = 11.6 mag star V* NN Lup. The we adapted the SAXO AO setup table in a 
We excluded two stars from the analysis, since the
data were collected on 2023-06-16, August 2024, meaning that currently the atmospheric conditions were very unstable during
under r­ elatively good atmospheric condi- faster correction mode is automatically the observing sequence.
tions. The difference in the AO correction selected up to G = 11.5 mag (in the past
between the two different AO modes the limit was 10.5 mag). This was actually

20 The Messenger 194 | 2025

Telescopes and Instrumentation DOI: 10.18727/0722-6691/5381

PoET: the Paranal solar ESPRESSO Telescope

Nuno C. Santos 1,2 6

Institute of Astrophysics and Space scopic Observations (ESPRESSO; Pepe et
Alexandre Cabral 3,4 Sciences, University of Coimbra, al., 2021), capable of achieving radial
Inês Leite 3,4 OGAUC Portugal velocity (RV) precisions down to 10 cm s–1
Alain Smette 5 7
Department of Physics, University — the typical amplitude of the signal
Manuel Abreu 3,4 of Coimbra, Portugal induced by an Earth-like planet orbiting
David Alves 3,4 8
Astronomy Department, University a Sun-like star — represent relevant steps
Jorge H. C. Martins 1 of Geneva, Switzerland in this effort. ESPRESSO will be comple-
Manuel Monteiro 1 mented by high-resolution optical and near-
André Silva 1,2 infrared spectrographs on ESO’s Extremely
Bachar Wehbe 3,4 The detection and characterisation of Large Telescope (ELT), for example with
Jimmy Arancibia 5 other ‘Earths’, orbiting other suns, is a the ArmazoNes high Dispersion Echelle
Gerardo Ávila 3 bold objective of present-day astro- Spectrograph (ANDES; Marconi et al.,
Stephane Brillant 5 physics. However, this quest is severely 2024), whose design is optimised for the
César Cárdenas 5 challenged by astrophysical ‘noise’ from detection of exoplanet atmospheres.
Ricardo Clara 3 the host stars, whose signatures distort These instruments will be key to following
Ricardo Gafeira 6,7 the observed spectra. Motivated by this up Earth-like planets detected by missions
Daniel Gaytan 5 problem, we are building a dedicated such as ESA’s PLAnetary Transits and
Christophe Lovis 8 facility, the Paranal solar ESPRESSO Oscillations (PLATO; Rauer et al., 2024).
Nicolas Miranda 5 Telescope (PoET). PoET will collect solar
Pedro Moreno 1 light and channel it into the ESPRESSO
António Oliveira 3,4 spectrograph, allowing us to use the Sun The stellar challenge
Angel Otarola 5 as a proxy to unambiguously identify
Francesco Pepe 8 and understand the sources of relevant In this quest, the greatest challenge to
Pascual Rojas 5 variability in solar-type stars. overcome is related to stellar physics, or
Ricardo Schmutzer 5 rather to the astrophysical ‘noise’ coming
Danuta Sosnowska 8 from the host stars that distorts the
Pierre van der Heyden 5 The quest for other Earths observed spectra.
Khaled Al Moulla 1
Vardan Adibekyan 1 More than 5000 extrasolar planets have The physical processes underlying mag-
Alba Barka 1,2 been confirmed to date. Present-day netically active features (spots, faculae;
Susana C. C. Barros 1,2 ­discoveries have already allowed one for example, Shapiro et al., 2016) produce
Pedro Branco 1,2 major conclusion: rocky planets seem to variations in the observed line profiles
Eduardo Cristo 1 be ubiquitous around solar-type stars. and positions. As activity-related features
Yuri Damasceno 1,2 Despite the impressive results, no Earth-­ appear and disappear from the stellar
Olivier Demangeon 1 analogue orbiting a Sun-like star has yet disc, they induce spectral variability on
William Dethier 1 been unambiguously discovered and timescales typical of the rotational period
João P. Faria 8 characterised, even if rocky planets around of the star, as well as of its long-term
João Gomes da Silva 1 lower-mass/smaller M-dwarf stars are magnetic cycle. Amplitudes can be as
Eduardo Gonçalves 1,2 within reach (for example, Faria et al., 2022; high as several tens or even hundreds of
Jennifer P. Lucero 1,2 Demangeon et al., 2021). metres per second in RV (for example,
José Rodrigues 1,2 Meunier et al., 2017; Faria et al., 2020).
Carmen San Nicolas Martinez 1,2 One of the main battle-horses for the Stellar granulation (Dravins, 1982) and
Ângela Santos 1,2 detection and characterisation of p-mode oscillations (for example, Chaplin
Sérgio Sousa 1 exoplanets is high-resolution spectros- & Miglio, 2013) are also relevant sources
Pedro T. P. Viana 1,2 copy (for example, Mayor, Lovis & Santos, of variability, inducing signals of up to
2014). Doppler spectroscopy measure- several metres per second, depending on
ments, complemented with transit photom- the spectral type (Dumusque et al., 2011).
1
Institute of Astrophysics and Space etry of planets transiting bright nearby For solar-type stars, granulation signals
Sciences, CAUP, Porto, Portugal stars (Lissauer et al., 2014), are particu- have timescales from a few hours to
2
Department of Physics and Astronomy, larly relevant and allow both their mass days, while for p-modes timescales are
Faculty of Sciences, University of and radius to be derived, and thus their of the order of a few minutes. An example
Porto, Portugal mean density. Complementary observa- of the impact of these processes in the
3
Institute of Astrophysics and Space tions and modelling allow their interior and RV of the Sun is shown in Figure 1.
Science, Faculty of Sciences, University atmospheres to be probed (for example,
of Lisbon, Portugal Ehrenreich et al., 2020). Although different approaches are pres-
4
Department of Physics, Faculty of ently used to model the signals produced
Sciences, University of Lisbon, Ground-based high-resolution spectro- by these different phenomena, none has
Portugal graphs such as the Echelle SPectrograph proven to correct RV time-series down to
5
ESO for Rocky Exoplanet and Stable Spectro- the required precision level (for example,

The Messenger 194 | 2025 21

Telescopes and Instrumentation Santos, N. C., et al., PoET: the Paranal solar ESPRESSO Telescope

Haywood et al., 2022). Furthermore, our as: if Earth were an exoplanet, would we star: only disc integrated spectra are
incomplete knowledge of stellar physics be able to tell whether it was habitable by obtained. This precludes a detailed analysis
can also severely impact our ability to observing its atmosphere? of the individual stellar features responsible
detect and characterise exoplanet atmos- for the observed spectral deformations.
pheres, or even produce systematics that
are several times stronger than the ones The solar promise
produced by the planetary atmosphere The Paranal solar ESPRESSO Telescope
itself (for example, Casasayas-Barris et The exoplanet community has recog-
al., 2021; Dethier & Tessore, 2024; and nised that progress in this field requires To find new answers we need to obtain
see Figure 1). The effects of these phenom- identifying in detail the different physical disc-resolved, high-precision spectra of
ena on the measurements of transmis- processes that drive stellar variations. In the Sun. Similarly to the best instrumenta-
sion spectroscopy can be particularly this context, the Sun is seen as the ideal tion used in exoplanet research facilities,
problematic, and raise questions such target and proxy: it is the only star we an adequate instrument has to offer a)
can resolve. Dedicated instruments have spatially resolved spectroscopy with very
been built, attached to high-precision high wavelength stability, b) very high
spectrographs, to observe the ‘Sun-as- spectral resolution (R = λ/Δλ ~ 200 000),
Figure 2. Left: Image of the Sun with Earth superim- a-star’ (for example, Zhao et al., 2023). to adequately resolve photospheric line
posed on the same scale. Upper-right: Radial-velocity Good examples are the High Accuracy asymmetries, and c) extended wavelength
time-series of the Sun (green points), obtained with
the HARPS-N solar telescope (Dumusque et al.,
Radial velocity Planet Searcher North coverage, for the simultaneous observa-
2021), when compared with the expected signature (HARPS-N) solar telescope (Dumusque tion of thousands of spectral lines prob-
of an Earth-like planet orbiting a Sun-like star (red et al., 2021) and its counterpart at the ing different physical conditions. This can
curve). Lower-right: simulated spurious signature of HARPS spectrograph (HELIOS). Overall, be achieved if we link the ESPRESSO
an absorption spectrum of an atmosphere-less
Earth-like planet centered in the sodium line (blue
these experiments have shed new light spectrograph to a solar telescope: the
curve), as caused by unaccounted center-to-limb on the problem, but also show that our Paranal solar Espresso Telescope (PoET1;
variations. The observed signal has a magnitude level current understanding is still inadequate see Figure 2).
similar to the expected absorption signature of the (for example, Milbourne et al., 2021).
atmosphere of an Earth-like planet. These plots illus-
trate the challenges for the detection and characteri-
In a nutshell (see also Leite et al., 2024),
sation of Earth-like planets orbiting other suns. The major drawback of these approaches PoET will consist of a 600-mm-diameter
Image sources: NASA SDO and ESO/M. Kornmesser. is the fact that the Sun is observed as a telescope designed to point to any resolved

4
Radial velocity (ms –1)

–2

–4

–6 HARPS-N observations
Sinusoidal curve (K = 0.08 ms –1, P = 365 days)

57400 57600 57800 58000 58200

Date (JDB-2450000)

1.0

0.8
Signal (ppm)

0.6

0.4

0.2

5895.4 5895.6 5895.8 5896.0 5896.2 5896.4

Wavelength in planet rest frame (Å)

22 The Messenger 194 | 2025

region in the solar disc and inject the light
into ESPRESSO using optical fibres. It ESPRESSO
will observe the Sun on different spatial
scales: from 55 arcsec in angular diameter,
the typical size of a medium-sized sunspot, Fibre cables
down to 1 arcsec, the typical spatial VLTI building from PoET
to ESPRESSO
scale of one solar granule. Simultaneous
full disc integrated observations (‘the-Sun-
as-a-star’) will also be possible using a H0
piggyback pointing telescope.

Instrument concept

To achieve simultaneous disc-integrated

and (arcsecond-level) disc-resolved
observations of the Sun, PoET will consist Paranal
of a three-telescope system, as shown by Observatory
the simple schematics in Figure 3.
The PoET frontends Figure 2. Left: Illustration of the location of PoET in
Paranal. Upper-right: Schematic of science fibres
The main telescope (MT) is from Officina
path from PoET to ESPRESSO (one floor below).
Stellare and its objective is the observa- As depicted in Figure 4 (right), in the Lower-Right: Dome concept with main telescope.
tion of small areas of the solar disc. It frontend of the MT, the f/12 beam passes
has a Gregorian configuration, chosen through a 90:10 beam-splitter and is
because of its intermediate focus. This directed to the aperture selector by a fold- ESPRESSO. A second refractor, with a
format, a standard for solar observations, ing mirror. The set of different apertures are 50-mm-diameter achromat with a focal
enables the introduction of a heat rejector positioned in the focal plane where light length of 350 mm, images the full Sun in
in the intermediate focal plane, rejecting is fed into small fibre patches, either the pointing camera sensor. The guiding
all the light that falls outside of a 4-­arcmin-­ directly (apertures from 1 to 5 arcseconds), and pointing cameras will be calibrated to
diameter field. This allows heat to be through demagnifying lenses (from the allow the identification of the area to be
reduced to less than 2% at the level of remaining apertures) or, alternatively, using observed by the MT.
the frontend focal plane. an integrating sphere (for the 55-arcsecond
mode). The selection is made using a
Although seeing-limited, one of PoET’s translation stage. The light from the exit Link to ESPRESSO
scientific objectives is to perform obser- of the selected aperture is injected into a
vations with a resolution on the order double scrambler to guarantee a homoge- The instrument will be installed behind
of an arcsecond, seeing allowing. The neous distribution into the science fibre the Very Large Telescope Interferometer
selected science apertures are 1, 2, 5, that connects with ESPRESSO. (VLTI) building and after the ‘no-build
10, 16, 30 and 55 arcseconds. For refer- zone’ (Figure 2). This location was chosen
ence, 16 arcseconds corresponds to the The apertures are metallic pinholes that as the preferred site for the instrument
angular size of Earth as seen from the reflect part of the incident light back to the based on its relative proximity to ESPRESSO
distance of the Sun. Owing to their small beam-splitter and redirect it to a guiding and relatively low wind speeds: to the
physical dimensions (from 35 μm to 2 mm camera (imaging system and sensor), north, the predominant wind direction, it
in diameter) and in accordance with where an image of the observed Sun is ‘protected’ by the VLTI building, which
our radiometric model, the selected tele- (with a 3-arcmin-diameter field of view) is avoids issues related to vibrations of the
scope aperture is required to ensure registered. This image will indicate the telescope structure.
enough light reaches ESPRESSO in all location of the observed region and can
configurations. be seen with either a 1-nm FWHM Hα The concept for the telescope dome is a
­filter or a 10-nm red filter (centred on the ‘sliding roof’ design. The roof consists of
Two simple refractors, the ‘science’ and same region). two sections that open in approximately
the ‘imaging’ telescopes, known jointly as the west–east direction. The telescope
the pointing telescope (PT), are piggy- The pointing telescope is comparatively will be hosted in a 5 m × 5 m space. An
backed onto the MT. Light from both the much simpler in design (Figure 4, left). additional room will be used to house the
MT and the PT will be collected by the A science refractor telescope (achromat electronics cabinets, computers, and
respective frontends and, via fibre links, lens), with an aperture of 60 mm and focal other auxiliary items.
delivered to the ESPRESSO spectrograph length of 100 mm, will image the Sun into
via its calibration unit. a 10-mm-diameter integrating sphere Science fibres, communication, and power
that injects the light into the ‘Sun’s disc-­ cables will be routed into an unused VLT
integrated science fibre’, connecting to Auxiliary Telescope hatch (H0), allowing

The Messenger 194 | 2025 23

Telescopes and Instrumentation Santos, N. C., et al., PoET: the Paranal solar ESPRESSO Telescope

Figure 3. Concept of PoET–ESPRESSO observations will be

PoET (left) and scheme
executed automatically.
of the main telescope
(right) with the position-
ing of the two frontends Observations will be made in both high-­
Pointing and the SHABAR seeing resolution and ultra-high-resolution modes,
telescope measurement device.
at spectral resolutions of ~140 000 and
~200 000, respectively (Pepe et al.,
SHABAR 2021). The baseline exposure time used
Main to define the instrumental design is
telescope 30 seconds, a compromise between the
Heat interest in resolving the solar oscillation
rejector signals and the overhead time related to
the readout of the ESPRESSO detector.

Several observing modes were defined,

Pointing
leveraging the optical setup of the PoET–
telescop(s)
ESPRESSO interface and PoET’s scientific
& SHABAR Main
requirements. The operation modes can
front
be divided into the following categories:
end
1. O
 bservations of the resolved Sun:
Main Pointing in this mode, PoET will acquire high-­
front end front end resolution spectra of a selected
Fibres to ESPRESSO (resolved) region of the solar disc using
Sun resolved and integrated the MT;
2. R
 esolved solar disc observations with
simultaneous ‘Sun-as-a-star’: in this
easier access for the fibres towards seeing during the day, using the Sun as mode, both the PT and the MT will be
ESPRESSO through the VLTI tunnel, with the source of light. For this purpose, injecting light into ESPRESSO;
minimal structural impact. Based on the we built a seeing measurement device, 3. Sun-as-a-star observations with simul-
observatory architectural drawings, we following the same concept as Sliepen et taneous wavelength calibration: meant
project around 80 m of cable for each al. (2010), to be used along with PoET. to acquire high-precision radial veloci-
science fibre, needed to deliver light Our SHABAR (Wehbe et al., 2024) is cur- ties time-series from disc-integrated
­collected by PoET towards ESPRESSO’s rently commissioned and being tested to exposures (‘Sun-as-a-star), mimicking
calibration unit, where PoET fibres will be ready for PoET’s first light. the usual stellar observations of
be connected. Inside the dome, the sci- ESPRESSO.
ence fibres are split via a set of collimators
that allow the insertion of a blue filter. Observations, operations and data In all cases, disc-integrated observations
Depending on the scientific requirement, will feed ESPRESSO’s fibre A, while disc
this filter can be selected to compensate As a visitor telescope at ESO’s La Silla– resolved observations will feed fibre B. As
for the signal loss in the blue region of Paranal Observatory, PoET is expected a consequence, it will not be possible to
the spectra, mainly caused by the fibre to be fully autonomous and not require obtain simultaneous wavelength calibra-
attenuation. any local intervention. As such, PoET’s tions while observing in the disc-resolved
software will oversee all operations of the mode. This will not impact the science
telescope and allow it to run either in a goals of PoET, as ESPRESSO drifts are a
SHABAR fully autonomous mode or managed few orders of magnitude smaller than the
remotely from our premises in Portugal. local velocities of the solar disc (for exam-
The MT will receive support from a cus- Typically, after the daily calibrations, ple, due to granulation).
tom instrument (known as SHABAR) to ­science operations will run a short script
measure the daylight seeing conditions. on the ESPRESSO workstation to pre- PoET data will be reduced with a new
Seeing is the effect of random fluctuations pare the instrument to execute PoET– version of the standard ESPRESSO Data
in the index of refraction throughout ESPRESSO observations. On its side, the Reduction Software (DRS). The DRS
Earth’s atmosphere, resulting in random PoET software will create science Obser- will produce the usual ESPRESSO sci-
fluctuations in the direction of light from a vation Blocks (OBs) and ingest them into ence-grade data products for the differ-
distant source. Scintillation is the random the ESPRESSO execution sequence ent observing modes detailed above.
fluctuation of the light intensity received. through the P2 Application Programming The reduced data will then be comple-
A correlation was found between the two. Interface2. From then on, and until sci- mented with an auxiliary FITS file contain-
With this concept, and making use of a ence operations require ESPRESSO for ing context images from the PoET cam-
non-telescopic method, we can measure the preparation of the night operations, eras as well as other relevant information

24 The Messenger 194 | 2025

 Pointing frontend Figure 4. Concept of the
Main frontend
two frontends and the
fibre injection system
Double scrambler Aperture selector and its aperture selector.

Guiding camera
with Hα and ND filters
Sun resolved z
fibre to ESPRESSO
Side View y

Integrated Sun
fibre to ESPRESSO
Aperture dimensions
Side View 1 arcsec / 35 µm
z
2 arcsec / 70 µm
y
5 arcsec / 175 µm

Selection direction
10 arcsec / 350 µm
16 arcsec / 560 µm
29 arcsec / 1.0 µm
55 arcsec / 1.9 mm
(two modes)

Aperture selector
x
100 mm Top View
(only selector) y

(for example, the seeing measured with expected to contribute to other science Dravins, D. 1982, ARA&A, 20, 61
Dumusque, X. et al. 2011, A&A, 525, A140
SHABAR). All data products will be cases, including solar and stellar physics.
Dumusque, X. et al. 2021, A&A, 648, A103
uploaded to the ESO Phase 3 archive. Ehrenreich, D. et al. 2020, Nature, 580, 597
Faria, J. P. et al. 2020, A&A, 635, A13
Acknowledgements Faria, J. P. et al. 2022, A&A, 658, A115
Haywood, R. D. et al. 2022, ApJ, 935, 6
Timeline and expected science impact We would like to acknowledge the fruitful discussions Leite, I. et al. 2024, SPIE 13096, 74
with the scientific community at the PoET Work- Lissauer, J. et al. 2014, Nature, 513, 336
The detailed design of the PoET tele- shops 3 organised in 2023 and 2024, that helped to Marconi, A. et al. 2024, Proc. SPIE, 13096, 1309674
scope is now concluded. The procure- define the final design of the instrument and the Mayor, M., Lovis, C. & Santos, N. 2014, Nature,
planning of the scientific observations. We would like 513, 328
ment of the main components, telescope to express our gratitude to the staff at Officina Stellare Meunier, N. et al. 2017, A&A, 597, A52
and mount, as well as the dome, is also for their support in the development of the solar tele- Milbourne, T. W. et al. 2021, ApJ, 920, 21
in progress. We expect to install the tele- scope. The project is funded by the European Union Pepe, F. et al. 2021, A&A, 645, A96
scope on Paranal in the second semester (ERC, FIERCE, 101052347). Views and opinions Rauer, H. et al. 2024, submitted to Exp. Astron.;
expressed are, however, those of the author(s) only arXiv:2406.05447
of 2025 and start operations soon after. and do not necessarily reflect those of the European Shapiro, A. I. et al. 2016, A&A, 589, A46
Union or the European Research Council. Neither Sliepen, G. et al., 2010, Proc. SPIE, 7733, 77334L
The main scientific motivation behind the European Union nor the granting authority can Wehbé, B. et al., 2024, Proc. SPIE, 13096, 1309683
PoET is to tackle the problem of ‘stellar be held responsible for them. This work was sup- Zhao, L. L. et al. 2023, AJ, 166, 173
ported by the Fundação para a Ciência e a Tecnolo-
noise’ in high-resolution spectroscopic gia (FCT) through national funds by these grants:
observations, both in precise radial UIDB/04434/2020 DOI: 10.54499/UIDB/04434/2020, Links
­velocities and in transmission/emission UIDP/04434/2020 DOI: 10.54499/UIDP/04434/2020,
spectroscopy. This will be fundamental and the PhD grant UI/BD/152077/2021 DOI: 1
 oET web page: https://poet.iastro.pt
P
10.54499/UI/BD/152077/2021. 2
ESO Phase 2 API: https://eso.org/sci/observing/
to the success of present and future
phase2/p2intro/Phase2API.html
efforts in exoplanet research, including 3
PoET workshops: https://poet.iastro.pt/events/
those linked with ELT instrumentation References
(for example, ANDES) and ESA missions
Casasayas-Barris, N. et al. 2021, A&A, 647, A26
(for example, PLATO). Moreover, PoET Chaplin, W. J. & Miglio, A. 2013, ARA&A, 51, 353
is presently raising interest among several Demangeon, O. et al. 2021, A&A 653, A41
scientific communities, and its data are Dethier, W. & Tessore, B. 2024, A&A, 688, L30

The Messenger 194 | 2025 25

Telescopes and Instrumentation DOI: 10.18727/0722-6691/5382

Flux Calibration for VLT and ELT Spectrographs

Sabine Moehler 1 be measured or modelled. The latter ratio between reference spectrum and
Wolfram Freudling 1 requirement demands an accurate char- observed data, as a result of different
acterisation of the atmospheric conditions resolution, imperfect wavelength calibra-
during a particular observing night. tion or radial velocity differences.
1
ESO
For many science cases, such as radial-­ Traditionally, reference data with a variety
velocity or equivalent-width measure- of resolutions and accuracy, derived from
ESO offers a range of optical and infra- ments, an absolute flux calibration is not ground-based or space-based observa-
red spectrographs on its telescopes, necessary, and a relative flux calibration tions for various spectral types, have
and more will be available once ESO’s is sufficient to reach the science goals. been employed for flux calibration of VLT
Extremely Large Telescope (ELT) starts Relative flux calibration corrects only the spectrographs. The most widely used
operating. Some, but not all, science instrumental signatures in the spectra, catalogues for optical spectrographs are
cases require the flux of spectra to be such as for instance the spectral energy those from Hamuy et al. (1994) and the
calibrated in relative or absolute units. distribution (SED) of the flat field in the CALSPEC database of the Hubble Space
The achievable accuracy of such a flux case of echelle data. Such a calibration Telescope (Bohlin, Gordon & Tremblay,
calibration differs with circumstances results in spectra that are free from small- 20141; most of which are too bright and/
and instrument. In this article, we pro- scale artefacts caused by the instrument or too northern for the VLT spectrographs).
vide an overview of the methods and such as wiggles and bumps, but the The data from Oke (1990) have systematic
routinely obtained accuracy for current overall shape of the continuum may not problems (as described in the original
Very Large Telescope and future ELT accurately represent the physical flux dis- publication) and are therefore no longer
spectrographs. tribution of the observed object. Reasons used for any VLT instruments.
for discrepancies between the observed
and the true shape include unaccounted For observations in the infrared wave-
Introduction slit losses, differences between the length range spectrophotometric stand-
assumed and the actual extinction, and ards that are used for optical wavelength
Spectroscopic flux calibration means variations in extinction or instrument ranges are not necessarily suitable,
the conversion of a spectrum from the response between the observations of because their spectra are usually known
measured signal in e – s –1 to physical flux the standard star and the science target. only up to about 1000 nm, and redder
units (erg cm –2 s –1 Å –1 or W m –2 nm –1). parts could contain spectral features or
The wavelength dependent conversion is In this paper, we provide an overview additional flux that affect the response
usually derived from the observations of of the methods and accuracy of spectral determination. The VLT Imager and
spectrophotometric standard stars, whose flux calibration of currently active Very Spectrometer for mid-InfraRed (VISIR)
absolute flux distributions are well known. Large Telescope (VLT) spectroscopic uses as reference data spectral templates
A ‘response curve’ is determined by taking instruments. for late-type stars from Cohen et al. (1999)
the ratio of the absolute flux distribution and the upgraded CRyogenic high-­
and the observed spectrum. Reduced resolution InfraRed Echelle Spectrograph
science spectra are then multiplied by VLT spectroscopic standard stars (CRIRES+) uses model spectra of optical
this response curve to flux-calibrate them. spectrophotometric standard stars of
Spectroscopic flux calibration relies on spectral types B4-A9.
There are two types of flux calibration, the availability of sufficiently bright spec-
absolute and relative. Absolute flux cali- troscopic standards with known absolute For medium-resolution spectrographs
bration, by which the observed flux of a flux. These standards are then regularly covering wavelengths from the ultraviolet
spectrum is converted to absolute flux observed to calibrate the combined to the near-infrared like X-shooter, accu-
units, is comparable to the photometric throughput of the atmosphere, telescope rate photometric reference data are diffi-
calibration of images. Absolutely flux-­ and instrument optics. From these obser- cult to obtain. In such cases, stellar model
calibrated spectra represent the correct vations, a response curve can be derived. spectra of hot white dwarfs can be used
flux distribution of the object, both in value The response curve can then be used to for flux calibration. The advantages of
and shape, and can thus be compared convert, for a given airmass and instru- using model spectra (which are also avail-
directly to independently observed pho- ment setup, the photon counts of a raw able for some of the CALSPEC stars) are
tometry or model spectra. Absolute flux spectrum to physical units. The response high spectral resolution and the absence
calibration requires that the total flux that needs to be tracked because of long- of noise and atmospheric absorption fea-
enters the telescope, for both the stand- term changes in the instrument and tele- tures like telluric lines. Model spectra can
ard star and the science target, can be scope (for example, mirror reflectivity), as be determined to a high level of accuracy
determined, and that the wavelength-­ well as changing atmospheric conditions for hot white dwarfs, which have relatively
dependent extinction for both a flux cali- (for example, CO2 or ozone abundance). simple atmospheres, especially when
brator and the science observations can Ideally the spectrum of a spectrophoto- compared to cool main sequence stars.
be determined. The former requirement metric standard star contains neither From a practical point of view their spectra
is difficult to satisfy since flux lost by the emission nor absorption lines, because offer another advantage in that they have
finite slit width or fibre diameter needs to such features can cause residuals in the a small number of smooth and wide

26 The Messenger 194 | 2025

 75° Figure 1. Distribution of the X-shooter spectrophoto-
60° metric standard stars on the sky. The two northern
stars are the HST CALSPEC standard stars GD71
45° and GD153.

30°
finely sampled reference data are
Declination (deg)

15° required to catch the small-scale varia-

–150° –120° –90° –60° –30° 0° 30° 60° 90° 120° 150° tions in the instrument response.
0°
Observed spectra are affected by telluric
–15° absorption, whereas the physical stellar
model spectra used as a comparison are
–30° not. In order to maximise the wavelength
range usable to compute the response
–45° curve, the observed spectra are corrected
–60° for telluric absorption which depends on
–75° Right ascension (deg) the atmospheric condition at the time of
observation. In the X-shooter pipeline this
correction is performed by selecting the
absorption lines. This limits the impact of slope of flux calibration using these best-fitting telluric model spectra from
small deficiencies in the wavelength cali- model spectra is accurate to about 5% a library of model spectra (Kausch et al.,
bration or differences in resolution com- (Sana et al., 2024) 2015). For UVES spectra, wavelength
pared to the case of cool main sequence regions with strong telluric absorption
star spectra with many more and much are excluded from the fit in the standard
narrower lines. Calibration of VLT spectrographs settings of 760 nm and 860 nm.

The finely sampled noise-free model ESO operates a range of spectrographs FORS2, with its lower resolution, has so
spectra can be interpolated to the that serve different science use cases far relied primarily on the standard stars
observed wavelength grid, so that the and employ a range of different techniques of Hamuy et al. (1994). The spectra were
ratio of the reference and observed spec- to produce data of different natures. masked to avoid telluric and stellar lines,
tra can be determined at the full instru- Accordingly, the flux calibration for each which in some cases resulted in a large
mental resolution instead of having to of them follows a specific plan. This cali- fraction of the spectrum being masked.
integrate the flux to the large bin size of bration plan for each instrument is availa- The FORS pipeline is currently being
low-resolution empirical reference data. ble at the instrument web pages2. updated to use the same standard stars
This enables the fitting of small-scale and methods as UVES and X-shooter.
instrumental variations that were previously In the case of FORS, the standard-star
lost. X-shooter was the first VLT instru- Long-slit spectrographs model spectra are convolved to lower
ment for which this approach was imple- spectral resolution to avoid large residuals.
mented consistently. The sample of The three optical/NIR slit spectro- The telluric correction is carried out
X-shooter flux standard stars consist of graphs currently operated at the VLT by directly fitting a physical model to the
six hot, white, hydrogen-atmosphere (the FOcal Reducer and low dispersion observed spectra (Smette et al., 2015;
white dwarfs (two of which are also part Spectrograph 2 [FORS2], the Ultraviolet Kausch et al., 2015). The latest release
of CALSPEC) and one hot, helium-rich and Visual Echelle Spectrograph [UVES], of the FORS pipeline contains these
pre-white dwarf. Figure 1 shows their dis- and X-shooter) perform only relative flux improvements.
tribution on the sky. Because the model calibration, because the science targets
spectra used as reference data do not are generally observed with narrow slits
contain telluric absorption, which becomes and the slit losses are hard to quantify Fibre spectrographs (ESPRESSO)
substantial in the near-infrared, a simple automatically. Readers interested in this
telluric correction was implemented. In issue are referred to Chen et al. (2014) The Echelle SPectrograph for Rocky
addition, fit points were defined to avoid and Manara et al. (2021). Exoplanet and Stable Spectroscopic
regions of very strong telluric absorption Observations (ESPRESSO) is a high-­
as well as the cores of the stellar lines, The experience with the X-shooter flux resolution spectrograph that uses fibres
where differences in resolution and/or calibration resulted in the switch of the instead of physical slits. Several fibres
radial velocity can create strong but narrow UVES reference data from the previous are used to form a ‘pseudo slit’. The total
residuals. The median values of the raw inhomogeneous collection to the X-shooter flux collected by a fibre depends on the
response over pre-­defined windows at stars in 2020. Figure 2 compares the exact two-dimensional positioning of the
the fit points are then used to fit the spectrum of the star κ Lep calibrated fibres relative to the source. It is therefore
response curve. More details can be using the old and new reference data even more challenging to quantify the
found in Moehler et al. (2014). The overall and pipelines 3. One can clearly see that ‘slit’ losses for a particular observation.

The Messenger 194 | 2025 27

Telescopes and Instrumentation Moehler, S., Freudling, W., Flux Calibration for VLT and ELT Spectrographs

1e–11 Figure 2. Comparison of the flux-calibrated spectrum for

8 the bright star κ Lep (observed 2019-01-17T01:01:23.629),
processed with the response curve derived from the
associated flux standard star EG21. The black spectrum
shows the result of the flux calibration using the old
7
UVES response determination, the red one the result
using the new UVES response determination (offset
for easier comparison).
6
Flux (10 –16 erg cm –2 s –1 Å –1)

5 The K-band Multi Object Spectrograph

(KMOS) also provides absolute flux cali-
bration but uses as reference data hard-
4
coded model spectra per spectral type
and magnitude, which consist of a Planck
3 blackbody curve with added absorption
lines, whose depth is adjusted to the
observed data. The final response curve
2 is thus a combination of instrumental
response and telluric absorption
1
The Enhanced Resolution Imager and
Spectrograph (ERIS-SPIFFIER) uses the
0 same standard stars as X-shooter and
3900 4000 4100 4200 4300 4400 4500 a similar procedure to determine the
Wavelength (Å) response curves.

ESPRESSO uses very low-resolution of the positioning of the sources relative Spectrographs without flux calibration
spectra of bright stars derived from to the detector. Slit losses can occur
ground-based observations (Hamuy et ­during the extraction of spectra for indi- The VLT spectrographs VISIR (slit) and
al., 1994), i.e., including telluric absorp- vidual targets, but these can in principle CRIRES+ (fibre) monitor the instrument
tion, as reference data. The pipeline be quantified to very high accuracy. The throughput/efficiency using standard
­corrects the observed flux for slit losses goal of the flux calibration of IFUs is to stars, but do not flux-calibrate the science
taking into account the seeing (described convert the total flux contained within data. For details see Table 1.
by a Moffat function) compared to the each wavelength plane of the data cube
diameter of the fibre. Slit losses become to physical units. For a multi-fibre instrument like GIRAFFE
significant if the seeing disc is larger than the cost of observing time to obtain flux
the fibre. The effects of positioning uncer- The Multi Unit Spectroscopic Explorer calibration was considered too high. The
tainty are not considered. The pipeline (MUSE) provides absolute flux calibration throughput of the instrument is monitored
computes the absolute efficiency at refer- and uses a mixed set of reference spectra, only in the IFU (ARGUS) mode.
ence wavelengths by comparing it to the some of them high-resolution model
reference spectrum. The result is then spectra and some data observed from The ERIS Near Infrared Camera System
interpolated to the ESPRESSO wave- space with lower resolution, all without (ERIS-NIX) instead does not monitor the
length scale using cubic splines. The pre- telluric absorption. The MUSE resolution efficiency of its long-slit mode because
cision of the flux calibration is expected is sufficiently low to cause problems if there are no suitable flux standard stars
to be low because of the highly variable high-resolution model spectra are used for the L and M bands.
fibre losses. as reference data without convolving
them first to a lower resolution, as for The Spectro-Polarimetric High-contrast
FORS2. The pipeline determines a first Exoplanet REsearch instrument Integral
Integral field spectrographs (MUSE, response curve by comparing the Field Spectrograph (SPHERE-IFS) observes
KMOS, ERIS-SPIFFIER) observed spectrum to the reference one. some of the CRIRES flux standard stars,
Next it interpolates across regions of but the data are not processed.
Integral field spectrographs (IFUs) pro- ­telluric absorption and determines a telluric
duce full position–wavelength cubes of a correction spectrum from the normalised
field using an image slicer that cuts the spectrum, assuming that the standard ELT spectrographs
field of view into slices that are then fed to star spectrum is smooth across the telluric
spectrographs. They do not suffer from regions. The final response curve is then All Extremely Large Telescope (ELT)
slit losses during the observation since linearly extrapolated to the largest possi- instruments will use adaptive optics,
the flux is collected mostly independent ble wavelength range and smoothed. which can cause variable flux losses

28 The Messenger 194 | 2025

 Wavelength Telluric Source Accuracy
coverage Flux correction for reference Type of of flux
Instrument Mode (microns) Resolution calibration of flux std data reference data calibration
No, science goals do not require it – this Model spectra
43000 / instrument is designed to measure radial from T. Rauch
CRIRES+ Slit 0.95–5.3 No Model N/A
86000 velocities and weak lines; efficiency & P. Coelho
monitoring only (priv. comm.)
No, no L or M band spectro-photometric
ERIS NIX Slit 3.05–4.05 900 N/A N/A N/A N/A
standards are available
5000 / Moehler
ERIS SPIFFIER IFU 1.09–2.47 Yes Yes Model Unknown
10000 et al. (2014)
70000– Hamuy et al.
ESPRESSO Fiber 0.38–0.79 Yes No Ground Unknown
190000 (1994) (HR stars)
ground Hamuy
Ground
260– et al. (1994);
FORS2 Slit 0.33–1.1 Yes No / model Unknown
5200 space/model
/ space
HST CALSPEC
No, because science cases do not
11000–
GIRAFFE Fiber 0.37–0.95 need it; observations of flux standards N/A N/A N/A N/A
39000
are not part of the calibration plan
11000– No, because science cases do not
GIRAFFE IFU 0.37–0.95 No N/A N/A N/A
39000 need it; efficiency monitoring only
10% on
Extended Model
continum
Hipparcos (blackbody
2000– ~15% wiggles
KMOS IFU 0.8–2.5 Yes No Compilation + absorption
4200 in case of
(XHIP) (Anderson lines, based on
good seeing
& Francis, 2012) spectral type)
observations
Continuum
Moehler
Model 5–10%, with
MUSE IFU 0.47–0.93 3000 Yes Yes et al. (2014);
/ space long-scale
CALSPEC (HST)
wiggles < 5%
No, because science casse do need
SPHERE IFU 0.95–1.65 30 / 50 it;observations of flux standards are N/A N/A N/A N/A
not part of the calibration plan
40000– Moehler
UVES Slit 0.3–1.1 Yes No Model Unknown
110000 et al. (2014)
No, because flux uncertainty
is dominated by the extremely high
350 / background, the sources are a percent Cohen
VISIR Slit 7.7–24.0 No Space N/A
25000 or less than the sky contribution and et al. (1999)
the Poisson errors form the background
is huge;efficiency monitoring only
Small-scale
wriggles
3200– Moehler about 2%;
X-shooter Slit 0.3–2.5 Yes Yes Model
18400 et al. (2014) slope about
5–10% (Sana
et al., 2024)

and therefore present special calibration 3–13 microns. It will use the same flux Table 1. An overview of the type of flux calibration
implemented at the VLT spectrographs.
challenges that are not addressed here. standard stars as VISIR.

The High Angular Resolution Monolithic The Multi-AO Imaging Camera for Deep
Optical and Near-infrared Integral field Observations (MICADO) offers medium-­ although with varying accuracy. A sum-
spectrograph (HARMONI) provides low- resolution IFU and slit spectroscopy at mary of the current state is given in
to medium-resolution IFU spectroscopy 0.8–2.5 microns. Suitable flux standard Table 1. The calibration plan for flux cali-
at 0.45–2.8 microns over a range of stars have been defined from X-shooter bration is the result of a careful instru-
resolving powers and will use the X-shooter observations of nearby hot white dwarfs. ment-specific cost-benefit analysis for
flux standard stars. each instrument, that might evolve over
the course of time. In some cases, it
The Mid-infrared ELT Imager and Spectro- Summary might be necessary for science pro-
graph (METIS) provides high-­resolution grammes to collect additional calibration
IFU spectroscopy at 3–5 microns and Spectrophotometry can be carried out data to improve on the flux calibration
low-resolution slit spectroscopy at with the majority of VLT spectrographs, that can be achieved with the standard

The Messenger 194 | 2025 29

 Telescopes and Instrumentation Moehler, S., Freudling, W., Flux Calibration for VLT and ELT Spectrographs

calibration plan. Any additional time References Links

needed for such extra nighttime calibra- 1
Anderson, E. & Francis, Ch. 2012, Ast. Lett., 38, 331 CALSPEC database: https://www.stsci.edu/hst/
tions must already be requested in Bohlin, R. C., Gordon, K. D. & Tremblay, P.-E. 2014, instrumentation/reference-data-for-calibration-
Phase 1. PASP, 126, 711 and-tools/astronomical-catalogs/calspec
Chen, Y.-P. et al. 2014, A&A, 565, A117 2
Paranal instrument details: http://www.eso.org/sci/
Cohen, M. et al. 1999, AJ, 117, 1864 facilities/paranal/instruments.html
Hamuy, M. et al. 1994, PASP, 106, 566 3
ESO data reduction pipelines: https://www.eso.
Acknowledgements
Kausch, W. et al. 2015, A&A, 576, A78 org/sci/software/pipe_aem_main.html
Manara, C. F. et al. 2021, A&A, 650, A196
We thank Lodovico Coccato, Mark Neeser, Isabelle
Moehler, S. et al. 2014, A&A, 568, A9
Percheron, and Valentin Ivanov for providing sub-
Oke, J. B. 1990, AJ, 99, 1621
stantial parts of the information presented here.
Sana H. et al. 2024, A&A, 688, A104
Smette, A. et al. 2015, A&A, 576, A77
Y. Beletsky (LCO)/ESO

Looking almost like a

watercolour painting, this
stunning photograph
of comet C/2024 G3
(ATLAS) was taken by
Yuri Beletsky on
19 January from ESO’s
Paranal Observatory in
Chile. The comet poses
next to one of the Auxiliary
Telescopes of ESO’s
Very Large Telescope
Interferometer.

30 The Messenger 194 | 2025

 ESO/VVVX survey

This image shows a detailed infrared view of

Messier 17, also known as the Omega Nebula or
Swan Nebula, a stellar nursery located about 5500
light-years away in the constellation Sagittarius.
This image is part of a record-breaking infrared
map of the Milky Way containing more than
1.5 billion objects. ESO’s VISTA (the Visible and
Infrared Survey Telescope for Astronomy) captured
the images with its infrared camera VIRCAM. The
data were gathered as part of the VISTA Variables
in the Vía Láctea (VVV) survey and its companion
project, the VVV eXtended survey (VVVX).

The Messenger 194 | 2025 31

Astronomical News

F. Millour/ESO

Apart from actually going into space, you probably

won’t feel much closer to the stars than at the
top of the staircase at ESO’s Paranal Observatory
shown in this picture. At 2635 metres (plus some
twenty steps) above sea level in the Atacama Desert
of Chile, you see a wealth of stars that remain hidden
in other places, because Paranal boasts the darkest
skies of all major observatories on Earth.

32 The Messenger 194 | 2025

 Astronomical News DOI: 10.18727/0722-6691/5383

Distributed Peer Review at ESO: Demonstrating Success

and Evolving Through Period 115

Tereza Jerabkova 1,2 fast-track channels and yearly cycles. Background and objectives
Ferdinando Patat 1 ESO is also exploring further innova-
Dario Dorigo 1 tions to optimise this process. The DPR process aligns with ESO’s com-
Fabio Sogni 1 mitment to fairness and efficiency in pro-
Francesca Primas 1 posal evaluations, introduced as a neces-
Annalisa De Cia 1 Introduction sary tool to manage the growing number
Elisabeth Renate Hoppe 1 of proposals that made it challenging for
The Distributed Peer Review (DPR) para- traditional panels to maintain high-quality
digm has emerged as a promising alter- reviews. At present, ESO assigns propos-
1
ESO native to traditional expert panel reviews, als requesting less than 16 hours of
2
Department of Theoretical Physics driven by the rapidly growing number of observing time — around 50% of all pro-
and Astrophysics, Faculty of Science, proposals submitted to large astronomical posals but only 20% of the allocated time
Masaryk University, Brno, facilities. By actively involving Principal — to DPR, while panels still oversee most
Czech Republic Investigators (PIs) and Co-Investigators of the time allocation. Potentially sensitive
(Co-Is) in reviewing one another’s pro­ cases, such as joint programs with ALMA,
posals, DPR seeks to distribute the work- XMM-Newton and other exceptions, are
ESO’s Distributed Peer Review (DPR) load more evenly while maintaining the reviewed by panels irrespective of the
has transformed proposal evaluations quality of evaluations. After initial deploy- time request, as outlined in ESO’s DPR
by fostering efficiency and community ments at Gemini Observatory, ESO con- guidelines1. The choice to maintain both
involvement, making it an essential ducted an early pilot project (Patat et al., DPR and traditional panels offers ESO
tool for handling the large volume of 2019), and the Atacama Large Millimeter/ valuable flexibility, allowing it to address
proposals traditional panels cannot submillimeter Array (ALMA) introduced specific proposal types that may require
review alone. A key strength of DPR is DPR for its Cycle 8 (although based on panel oversight in the future. This hybrid
its inclusion of the entire community, rankings, not grades as for ESO; Carpenter system, with DPR alleviating the workload
engaging researchers at all career levels et al., 2022). The first comprehensive on panels, ensures that panels can main-
and identifying expert reviewers. This assessment of ESO’s DPR outcomes was tain a high standard of review quality.
article summarises updated findings up presented by Jerabkova et al. (2023), lay- The DPR system leverages ESO’s User
to Period 115, focusing on expertise ing a solid foundation for broader adop- Portal, where each user is required to
assignment, DPR comment usefulness tion. In this article, we update and expand provide two to five keywords representing
and user satisfaction. DPR’s success upon these insights through Period 115. their scientific expertise. Similarly, pro-
supports its planned expansion into posals include selected keywords from
the same pool, which are used to calcu-
P110 Histogram P115 Histogram
late expertise-­match scores between
reviewers and proposals. These scores
1750
Random assignment Random assignment are central to the proposal distribution
2000 ∑iscore i = 1855.05 ∑iscore i = 1575.37
\# of 0 = 2252
1500
\# of 0 = 1761
process. Each submitting team, repre-
1250 sented by a PI, must nominate one reviewer
1500 DPR Assignment DPR Assignment
who is responsible for evaluating DPR-­
Frequency

Frequency

∑iscore i = 6226.15 ∑iscore i = 4912.59

1000
\# of 0 = 60 \# of 0 = 47 assigned proposals from their peers.
1000 \# of proposals submitted = 435 750 \# of proposals submitted = 344

500
500
250
Figure 1. Expert reviewer assignments in the DPR for
0 0
0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 Period 110 (left) and Period 115 (right). The bottom
Keyword match score (P110) Keyword match score (P115) panels present the assignment matrix, in which DPR
proposals are on the horizontal axis and DPR reviewers
P110 Matrix plot P115 Matrix plot on the vertical axis. Although the matrix is not
2.00 2.00
300 square, it is sorted so that the main diagonal corre-
350
1.75 1.75 sponds to each proposal’s delegated reviewer —
300 250 often the PI or a co-I. The colourmap shows keyword-­
1.50 1.50 based expertise scores: a high score along the
Reviewers (sorted)

Reviewers (sorted)

250 200 diagonal verifies that PIs or co-Is indeed have strong
1.25 1.25
expertise for their own proposals, serving as a self-­
Scores
Scores

200
1.00 150 1.00 consistency check.
150 0.75 0.75
100 In the top panels, the grey lines indicate how a purely
100 0.50 0.50 random assignment would appear. Because most
50 keyword-matched scores are zero, the random
50 0.25 0.25
assignment fails to align expertise with proposals. By
0 0.00 0 0.00 contrast, the final assignment (shown in blue) largely
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 avoids zero-score pairings, confirming the necessity
Proposals (sorted) Proposals (sorted) of a more sophisticated reviewer assignment process.

The Messenger 194 | 2025 33

Astronomical News Jerabkova, T., et al., Distributed Peer Review at ESO Update Through Period 115

Distribution of expertise scores vs self-identified expertise (P110-P115) Figure 2. Comparison between the reviewer-­
proposal match scores — calculated from specified
2.00 keywords — and reviewers’ self-assessed expertise.
This Figure demonstrates how closely the keyword-­
based matching aligns with the reviewers’ own percep-
1.75 tion of their expertise, serving as a validation for the
automated assignment approach.
1.50
Expertise score

1.25 machine learning approach for reviewer

assignments (Carpenter, Corvillón &
1.00 Shah, 2024).

0.75 By addressing the limitations of keyword-­

based assignments and leveraging
0.50 emerging technologies, DPR can continue
to evolve as a sustainable and fair peer-­
0.25 review method that adapts to an ever-­
Linear increasing volume of astronomical
0.00 trend line proposals.
Expert Intermediate Non-expert
Self-identified expertise
User satisfaction

Key findings and updates alongside rapidly advancing scientific Jerabkova et al. (2023) first examined
fields, mismatches can arise over time. user satisfaction data from Period 110,
2. U
 ser Inconsistencies: Some users fail finding that feedback under DPR was
Algorithm performance to update their keywords regularly, or generally better received than traditional
they assign them improperly, both of panel comments — especially for rejected
The assignment algorithm remains a criti- which can adversely affect the accuracy proposals, where constructive input is
cal component of DPR’s success. As of reviewer–proposal matches. critical. Subsequent user surveys are
illustrated in Figure 1, the final matching ­systematically run each Period. They are
of proposals to reviewers outperforms Figure 2 demonstrates the link between built into the DPR evaluation system
a purely random approach, resulting in an keyword-based match scores and and receive responses from typically 50%
optimal distribution of expertise across reviewers’ self-identified expertise. While of the PIs. The outcomes indicate that
the submitted proposals. This robust these results validate the general reliability since the implementation of DPR PIs with
method ensures the integrity of the review of using keyword vectors to define reviewer accepted proposals consistently rate DPR
process, even as the number of proposals expertise, they also underscore potential feedback as valuable, whereas rejected
and reviewers continues to grow. In our pitfalls arising from outdated or misapplied proposals attract more mixed responses.
previous work (Jerabkova et al., 2023), we keywords. Ongoing efforts by Amado Despite these variations, a large fraction
focused on Period 110 and partially on et al. (in preparation) aim to refine the of DPR users now report that comments
Period 111, establishing the first statistical keyword framework, making it more flexi-
analysis of the DPR’s performance. Build- ble and adaptive to scientific evolution.
ing on those findings, the current results In parallel, ESO is communicating closely Table 1. Summary for each period for both DPR and
reinforce how crucial it is to match each with ALMA, which has begun adopting a panels.
proposal with a suitably qualified reviewer.
Period Proposals Feedback Accepted Rejected %
DPR P110 435 1358 349 1009 31.22%
Advantages and challenges of keyword-­ Panel P110 429 124 50 74 28.90%
based assignments DPR P111 419 2708 1138 1570 64.63%
Panel P111 401 247 121 126 61.60%
Keywords continue to play a dual role in DPR P112 451 1911 732 1179 42.37%
facilitating expertise matching. On the Panel P112 442 222 113 109 50.23%
positive side, they are intuitive to set and DPR P113 402 2555 1239 1316 63.56%
interpret, enabling the community to self-­ Panel P113 378 206 112 94 54.50%
regulate how expertise is represented.
DPR P114 403 2413 1003 1410 59.91%
However, some challenges persist:
Panel P114 448 245 134 111 54.69%
1. S
 tatic Nature of Keywords: because DPR P115 344 1481 698 783 43.05%
keywords do not automatically evolve Panel P115 416 123 73 50 29.57%

34 The Messenger 194 | 2025

 Comment usefulness for DPR (Period 110–115) are mostly or fully useful, underscoring
the value of this more distributed approach
0.35 Accepted proposals fit
to peer review.
Rejected proposals fit
Overall In parallel, panel-based evaluations have
0.30 Accepted shown some improvement over the same
Rejected periods, likely thanks to the 50% reduction
in workload made possible by DPR. Histori-
Fraction of all comments

0.25
cally, users have viewed panel feedback
— especially for rejected proposals — as
0.20 less beneficial, a sentiment strongly
reflected in the Period 110 survey results.
However, current data (spanning Periods
0.15 110 to 115), as shown in Figures 3 and 4,
reveal a modest but encouraging upswing
in the proportion of panel comments
0.10
deemed mostly or fully useful, at least for
accepted proposals. While this positive
0.05 trend has not yet matched DPR levels, it
offers promise that, with continued effort
and streamlined workloads, panel com-
0.00 ments can further improve — particularly
Not useful Somewhat Mostly Fully where it matters most, i.e. for proposals
Comment quality that ultimately receive a rejection. Table 1
Figure 3. Histogram of user ratings for DPR feedback for accepted proposals, and the right panel (or second shows a number summary for each period
spanning Periods 110–115. The four categories — not subplot) shows ratings for rejected proposals. A notable for both DPR and panels, showing that the
useful, somewhat useful, mostly useful, fully useful — uptick in fully useful feedback is seen over time for both analysis is based on robust numbers.
reflect PIs’ perceptions of review comments they categories, suggesting that DPR has steadily improved
received. The left panel (or first subplot) shows ratings in delivering constructive and actionable comments.

Career stage and reviewer bias

Comment usefulness for panel (Period 110–115)

Accepted proposals fit Analyses of DPR grading trends reveal no

0.30 statistically significant bias in the scores
Rejected proposals fit
P110 rejected proposals trend
assigned by reviewers at different career
stages. Although there is a slight tendency
0.25 Overall
for senior scientists to award poorer
Accepted
grades, students continue to be a pivotal
Rejected part of the process, often providing some
Fraction of all comments

0.20 of the most constructive and well-received

comments. These observations are con-
sistent with earlier results (Jerabkova et
0.15 al., 2023), reinforcing the conclusion that
career-stage differences are not a major
driver of bias in the DPR framework.
0.10
An additional point worth highlighting is
the inclusivity of the DPR model, which
allows students and junior researchers
0.05
to serve as reviewers — an option not
typically available under traditional pan-
els. Interestingly, the feedback they sup-
0.00 ply is frequently rated as most useful by
Not useful Somewhat Mostly Fully
Comment quality
proposal authors (see Figure 5), validating
the merit of incorporating perspectives
Figure 4. The same as Figure 3, but illustrating user from reduced reviewer workloads, thanks to the from earlier-­career scientists. When stu-
ratings of panel reviews over Periods 110–115. While adoption of DPR, which allows panels to dedicate dents or junior researchers feel uncertain
user satisfaction with panel comments remains lower more time to each proposal. However, panel com-
overall compared to DPR, an upward trend in mostly ments for rejected proposals still show less favourable
about their evaluations, they may request
and fully useful ratings is observed — particularly for ratings, indicating the need for continued attention to permission to consult their supervisors,
accepted proposals. This improvement may stem feedback quality in these cases. who must abide by the same confidentiality

The Messenger 194 | 2025 35

 Astronomical News Jerabkova, T., et al., Distributed Peer Review at ESO Update Through Period 115

P (helpful comment | career stage) larger fraction of submissions to keep

the expert panel load at the current level.
While ALMA has implemented the DPR
model for 100% of their submitted pro-
Student 0.09 0.23 0.31 0.37 posals, ESO plans to retain its panels for
specific cases, such as joint ALMA–ESO
proposals and those with significant time
requests, ensuring a balanced approach
that combines DPR’s efficiency with
Career stage

the expertise of traditional committeesa.

Studies by Carpenter et al. (2022) on
Post-doc 0.11 0.22 0.32 0.35
ALMA’s implementation of DPR provide
valuable insights that support these
developments.

Expansion to fast-track channel

Professional 0.14 0.23 0.30 0.34
ESO is confident in deploying DPR for its
upcoming fast-track channel, as high-
lighted by Patat et al. (2024). This channel
Not useful Somewhat Mostly Fully is expected to streamline the evaluation
Review evaluation on short timescales while maintaining the
high standards of review quality estab-
Figure 5. Color-coded conditional probability map explore algorithms that ensure diversity lished by DPR.
showing how perceived comment usefulness varies
among the reviewers assigned to each
with reviewer career stage in the DPR. Each bin dis-
plays the probability that a reviewer at a given career proposal. Achieving this would require a
stage receives a specific usefulness rating from the significantly larger pool of reviewers to Conclusion
PIs, with values normalized across the usefulness balance these considerations effectively.
axis (i.e. each row sums to one). The data confirm
With DPR now operating across six
that comments from students are consistently rated
as highly useful, while senior researchers — though Implications and future directions ­periods, its benefits for the ESO commu-
often more stringent in grading — also provide valua- nity are clear. The process has reduced
ble, high-quality feedback. These results align with The successful integration of DPR into panel reviewer workloads, maintained
findings from Period 110, reinforcing the important
ESO’s evaluation process demonstrates fairness in proposal evaluations, and pro-
contribution of early-career scientists to the peer-­
review process. its viability as a scalable alternative to vided helpful feedback to PIs. The ongo-
­traditional panel reviews. The reduction ing refinement of algorithms and reviewer
in panel workload and the high satisfac- profiling ensures that DPR will continue
agreements. This process has been tion rates among users reinforce the to serve as a model for peer review inno-
invoked in several instances each semes- value of DPR. Nonetheless, there is room vation in astronomy and beyond.
ter, helping to ensure that reviewers at all for improvement:
career stages can participate confidently References
and responsibly. – Adaptive Algorithms: ESO is exploring
more sophisticated assignment algo- Carpenter, J. M. et al. 2022, PASP, 134, 045001
rithms, including those leveraging Carpenter, J. M., Corvillón, A. & Shah, N. B. 2024,
arXiv:2410.10009
Reviewer demographics and diversity machine learning, to further enhance Jerabkova, T. et al. 2023, The Messenger, 190, 63
the match between proposals and Patat, F. et al. 2019, The Messenger, 177, 3
One notable aspect of DPR is that the reviewers. Patat, F. et al. 2024, The Messenger, 193, 45
reviewer pool reflects the broader com- – Keyword Refinement: Efforts are under- Primas, F. et al. 2024, Proc. SPIE, 13098, 130980H
munity, with a slight overrepresentation way to make the keyword system more
of postdoctoral researchers. Unlike tradi- agile and reflective of contemporary Links
tional panels, ESO does not control the science.
1
composition of DPR reviewers in terms – Increased Feedback Participation:  SO’s DPR guidelines: https://www.eso.org/sci/
E
observing/phase1/distributed-peer-review.html
of gender, seniority or nationality. In con- Strategies to improve the response rate
trast, panel composition is carefully for feedback surveys are being devel-
curated to ensure diversity across these oped to expand the statistical basis for Notes
dimensions (Primas et al., 2024; Primas future analyses.
a
In the current implementation the proposals are split
et al., in preparation). While this organic
50/50 in number, while the time share is 80/20
representation of the community is valua- ESO plans to use DPR within its yearly (Panels/DPR).
ble, it may be interesting in the future to proposal cycle, likely covering an even

36 The Messenger 194 | 2025

 Astronomical News DOI: 10.18727/0722-6691/5384

ESO’s Scientific Visitor Programme in Garching

Michael Hilker 1 provide researchers with the opportunity and ­submit it via email to the Visitor
Harald Kuntschner 1 to engage with ESO’s cutting-­edge sci- ­Selection Committee (VSC) Chairperson
ence, science operations and technology at [email protected].
while contributing their expertise to foster
1
ESO innovation, promote scientific interaction Applicants are encouraged to coordinate
and advance ESO’s projects and goals. with an ESO staff member or Fellow
before applying so as to align their visit
The ESO-Garching Scientific Visitor with specific projects. Applications should
Programme invites astronomers and Eligibility criteria be submitted at least four months prior
researchers from around the world to to the planned visit to allow for practical
visit ESO in Garching for collaborative Eligible applicants for the regular visitor arrangements. The programme’s schedule
scientific activities. Open to scientists programme include scientists with a tends to be busiest during the summer
actively engaged in research, the pro- PhD in astronomy or related disciplines months (June–August), so visits outside
gramme pro­motes scientific interaction, actively engaged in astronomical this period are encouraged when possible.
innovation, and ESO’s role as a hub research. Early-career researchers (PhD
of astronomical excellence. Visitors students and postdocs up to three years
receive logistical and financial support, post-PhD) are encouraged to apply via Evaluation and selection
engage in ESO’s vibrant scientific life, the Early-Career Visitor Programme2.
and contribute to its projects. Applica- All applicants must have an active affilia- The VSC, comprising ESO Faculty
tions are reviewed by the Scientific tion and job contract during the visit. ­members and Fellow representatives with
­Visitor Selection Committee. This pro- diverse expertise, evaluates applications
gramme aligns with ESO’s mission While the programme is open to research- based on 1) scientific excellence, and 2)
to foster international collaboration in ers of all nationalities, preference is given potential contribution to ESO’s scientific
astronomy and its vision to work with to citizens of ESO Member States, the projects. The committee meets about
and for the astronomy community. host state Chile and ESO’s strategic every two months to review applications
­partner Australia. ESO is committed to and recommend candidates to the Head
diversity and inclusion, ensuring equal of the Office for Science.
Programme scope and objectives opportunities regardless of gender, age,
disability, ethnicity or other factors.
The Scientific Visitor Programme at Support and responsibilities of visitors
Garching1 supports long-term visits by
accomplished young and senior scientists, Application process Approved visitors receive a monthly allow-
enabling close collaboration with ESO ance for living expenses and family sup-
staff, fellows, and students. These visits, Scientists interested in applying should port, if applicable, reimbursement of travel
ranging from one month to a year (one to complete the application form (available costs, and accommodation in fully fur-
four months for Early Career Scientists2), through the Visitor Programme web page) nished ESO apartments, if available. They

Histogram of visitors with visitor months as line

30
Male Visitor months
Female
40
25
Count of male and female visitors

20
30
Visitor months

15

20

10
Figure 1. Number of male
and female scientific vis-
itors since 2001. The
5 10
green line gives the num-
ber of visitor months
accepted for each year
(y-axis on the right). The
0
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 effect of the pandemic in
Start year 2020/21 is clearly visible.

The Messenger 194 | 2025 37

Astronomical News Hilker, M., Kuntschner, H., ESO’s Scientific Visitor Programme in Garching

Percentage contribution by country IE BR Figure 2. Pie chart

SE 1.2% 1.5% showing the percentage
1.0% AR IN contributions of visitors
DK 1.2% 1.7% by country, with coun-
1.0% RU tries contributing less
PL 2.0% than 0.5% grouped into
1.0% Others. More than a
CA
Other MX quarter of all visitors
IT CH 2.2%
5.7% 1.0%
26.4% 0.7% between 2001 and 2024
JP CL
have Italian nationality.
1.0% 2.5%
BG Countries
1.0% Italy
Germany
United States
Spain
United Kingdom
France
Australia
Netherlands
Greece
Chile
Canada
Russia
India
Brazil
Argentina
Ireland
Bulgaria
Japan
Mexico
Poland
GR
2.7% Denmark
Sweden
NL
Switzerland
4.2%
Other
DE AU
10.4% 4.7%

FR
US 5.5%
8.5%
GB
ES 6.2%
6.5%

are provided with office space, a com- They have delivered lecture series, talks Historically the proportion of female
puter terminal, and administrative support. and seminars, worked on white papers, ­visitors was low (21.6% before 2022) but
and inspired ESO’s young fellows and has improved in the past three years,
During their stay, visitors are expected to students. These interactions have sparked reaching 38.6% thanks to increased
deliver a seminar or informal discussion, collaborations leading to successful awareness and training of the VSC and
engage with ESO’s scientific activities observing proposals and refereed articles. offered support.
and adhere to the ESO Code of Conduct.
Figures 1 and 2 illustrate the demographics
of scientific visitors since 2001. Over this References
Some statistics on past visitors period, the Office for Science in Garching ESO 1993, The Messenger, 72, 7
has hosted 402 visitors, with annual
For more than 30 years, ESO has been ­numbers ranging from 10 to 29, except
running a scientific visitor programme (see during the COVID-19 pandemic years. Visit Links
ESO, 1993). During this time, the Office for durations have spanned the full range 1
ESO’s Scientific Visitor Programme: https://www.eso.
Science in Garching has hosted hundreds supported by the programme, with an org/sci/activities/garching/personnelvisitors.html
of visitors, comprising a mix of experi- average length of 1.7 months. On average, 2
E arly-Career Scientific Visitor Programme:
enced, high-profile astronomers and early-­ about 30 visitor months are approved https://www.eso.org/sci/activities/garching/
personnelvisitors/Policy_Early-Career_Scientific_
career scientists. All have made significant each year, meaning that nearly three visi-
Visitors_Garching.html
contributions to ESO’s scientific activities. tors are present at ESO at any given time.

38 The Messenger 194 | 2025

Astronomical News DOI: 10.18727/0722-6691/5385

Usage of Artificial Intelligence by ESO Telescope Users

Pooneh Nazari 1 way that we approach research, coding, of AI also correlates with level of seniority,
Tereza Jerabkova 1 and potentially preparation for observing with 33% of PhD students, 27% of post-
Ferdinando Patat 1 proposals. Therefore, it is important for doctoral researchers, and 18% of faculties
ESO, as one of the forefront organisations using AI when writing proposals. This is
in building and operating ground-based illustrated in Figure 1b. Although a smaller
1
ESO telescopes, to investigate the use of portion of the faculty members seem to
LLMs in proposal preparation and review, use LLMs in proposal preparation, given
and, if necessary, establish policies their larger group size it still results in a
With the increasing integration of artifi- regarding this. In this article, we report larger number of faculty members using
cial intelligence (AI) and large language the results of a survey that we shared AI than postdoctoral researchers or PhD
models (LLMs) into various fields, it is with the ESO telescope users to shed students. Of those who use AI (about
crucial to understand their capabilities light on the current usage of AI in pro- 190 participants), most did not notice any
and usage within the scientific commu- posal preparation and review. This is part change in their success rate. More specifi-
nity. This study explores the adoption of the ongoing efforts at ESO (i.e., the cally, Figure 1c shows that only around
and impact of these technologies among STARS@ESOa working group; Jerabkova 10% of the participants who use AI across
astronomers using ESO telescopes, et al., 2024) to better understand the effect various career stages noticed a positive
specifically in the context of proposal of AI on proposal preparation and review. change in their success rate.
preparation and review. We shared a
survey with the ESO telescope users to We were also curious about how LLMs
investigate this further. We received Survey demographics and AI usage are used when drafting proposals. The
827 responses and found that around left panel of Figure 2 shows a summary of
20–30% of the participants use LLMs The survey on AI (in this context mostly what AI is used for. Across career levels,
when preparing proposals and about LLMs) usage by the ESO community in it is mostly used to enhance clarity and
3% of the participants use them when writing proposals was open for around readability. After that, it is mainly used for
reviewing proposals. We also found four and a half months. We contacted the title of the proposal. Finally, a minority
that there is a divide in how the usage around 2300 PIs who had submitted a use AI for fact checking, for calculations,
of AI is perceived when proposals are proposal in the last ten semesters and and/or for writing an entire abstract. Two
prepared/reviewed, pointing to the need got 827 responses (about 36%). Figure 1a participants mentioned that they use AI
for ESO to establish clearer guidelines. presents a summary of the career stages for writing the entire proposal including
These guidelines will be released for the of the survey participants. The collected suggestions for the science idea. When
next period. distribution and demographics are well it comes to reviewing proposals, the right
representative of the ESO community. panel of Figure 2 shows that most par­
Figure 1a also shows how much experi- ticipants (~97%) do not use AI for this
Motivation for the survey and its ence the participants have had of writing purpose. However, around 3% do use AI
outcome proposals. In brief, 55% of the survey when reviewing proposals, which accounts
participants have faculty positions, 30% for around 28 of the participants.
Artificial intelligence (AI) is used commonly are postdoctoral researchers, and 15% are
in all areas of science. A few years ago, PhD students. Faculty members mostly
the use of large language models (LLMs) have more than five years of experience Relevant comments
in problem-solving tasks became rela- and PhD students mostly have one to two
tively common by the release of ChatGPT years of experience of writing proposals. We also asked the participants to provide
and, following that, other similar conver- any relevant comments. From those who
sational models such as Google Gemini Of the survey participants only about provided them, two main general themes
and Claude. This has transformed the 23% use AI to draft proposals. The usage emerged. Around 80–90 participants
(~10%) expressed their concern regarding
the use of LLMs in either preparing the
a) b) c) proposals or reviewing them. Some went
1–2 No 80
8% No as far as suggesting banning its use
400 3–5 400 18% Yes Yes (positive)
>5
70 ­altogether or implementing a procedure
11%
60 to catch those who use LLMs for the
300 300
50
Number

8%
40
200 200 82% 27% 92%
30 89%
Figure 1. a) Number of participants in different
100 100 33% 20 92%
73% career stages; the colours present the number of
10 cycles for which they have written a proposal for
67%
0 0 0 ESO telescopes. b) The result of their answers to the
question: Do you use AI when preparing proposals?
Faculty

PhD

Faculty

PhD

Faculty

PhD
Postdoc

Postdoc

Postdoc

c) The result of their answers to the question: If you

use AI for proposals, have you noticed any change in
Career stage Career stage Career stage your success rate?

The Messenger 194 | 2025 39

Astronomical News Nazari, I., et al., Usage of Artificial Intelligence by ESO Telescope Users

­ roposal preparation and disqualify those

p What do you use AI for? Do you use AI to review proposals? Figure 2. Left: This plot
shows what AI is used for
who use it above a certain threshold. On 70
Clarity
Yes
Tite when preparing propos-
the other hand, some of those who were Fact check 3.4% als. Right: The portion of
60
concerned did not object to its use in Calculations the participants who use
50 Entire abstract
improving the language of the proposal, AI for reviewing proposals.

Number
Entire proposal
40
especially for those who are not native
speakers of English. Some of the con- 30 96.6%
cerns were related to the ethical aspect, 20
No
as the large language models (LLMs) are 10
trained on what is the intellectual prop- 0
Faculty Postdoc PhD
erty of others and therefore proper cita- Career stage
tions are required; the words “unethical”
and “plagiarism” were used in several
comments. Another concern that a few in favour of proposals written by LLMs the proposal preparation and review pro-
participants expressed was related to the (Jerabkova et al., 2024). Although some cess and, if so, to what extent they can
decline of creativity and the uniformity of participants in the other group were worried be employed. Currently, ESO only has a
all proposals when using such models. about declining creativity, one person in disclaimer at the start of the review pro-
Several participants also pointed to the this more positive group mentioned the cess such that the reviewers agree to not
need to double check what a LLM pro- use of AI for inspiration. share the proposals with third ­parties.
duces to avoid wrong statements. One With the implementation of the new
person suggested that a negative factor Other comments included suggestions ­Confidentiality Agreement for the Review
could be the use of AI to make the pro- to assess proposals in a different way if of the ESO Observing Proposals, stricter
posal sound interesting, which may LLMs are to be used more widely, for measures are now in place to safeguard
appeal to the younger reviewers, while example by putting less emphasis on sensitive information. Reviewers are
the proposal may have scientific issues. the clarity of the proposals and more on explicitly prohibited from using automated
A few people were also concerned about the science idea. A few participants processing methods, including AI tools
the energy consumption and carbon ­suggested asking for a disclaimer from such as ChatGPT, Google Gemini, or
dioxide emission when training the LLMs. proposers/reviewers to confirm whether Claude, to process, analyse or interpret
Some of those with concerns about the LLMs are used to produce/review pro- proposal content without prior written
use of AI in reviewing proposals particu- posals. Finally, we also received com- consent from ESO. This agreement
larly mentioned the issue of consent, ments on the shortcomings of our survey. ensures that confidential information
potentially hinting at the need to have the For example, some participants indicated remains secure and emphasises the
PIs’ agreement for their proposals to be that they use LLMs for coding and making importance of maintaining integrity and
uploaded into LLMs. Some were also plots, which was not given as an option authenticity in the review process. As we
suspicious that they received reviewer in the survey. As we wanted to keep the move forward in this rapidly evolving
comments generated by AI and that they survey short, we avoided asking more technological landscape, ESO acknowl-
were generic and not quantitative. One in-depth questions on the exact usage edges the need for further specific
person was worried that the distributed of LLMs when reviewing proposals. ­guidelines and policies to address the
peer review process might lose its mean- Therefore, a few participants elaborated ethical and practical challenges associ-
ing if AI is used in reviewing proposals. on their usage when reviewing the pro- ated with AI usage in proposal prepara-
posals. These included asking AI questions tion and review.
The second theme included participants on a subject to learn a topic fast enough
who were more positive about using LLMs to be able to assess the proposals fairly
in the preparation of proposals or review- and using LLMs to refine and increase Acknowledgements
ing them. This group consisted of around the clarity of their comments. We thank the ESO Director for Science for support-
60 participants (~7%) with roughly half of ing this research. Moreover, we thank all the ESO
them emphasising the potential of AI in telescope users who participated in the survey.
helping non-native speakers of English to Closing remarks PN acknowledges support from the ESO and IAU
Gruber Foundation Fellowship programmes.
produce higher quality text, which can
increase the fairness of the process. This To conclude, around 20–30% of the survey
more positive group also included partici- participants already use LLMs in the pro- References
pants who are thinking about starting posal preparation process and there
Jerabkova, T. et al. 2024, arXiv: 2407.02992
to use AI in proposal preparation and/or seems to be a divide between those who
review in the future. One person even strictly disagree with the usage of LLMs
suggested integrating it into the proposal in proposal preparation and those who Notes
tool. Another person noted that in the are more positive about it and see poten- a
 TARS stands for Scientific Text Analysis with
S
review process, LLMs might have fewer tial benefits. These findings indicate that RobotS
issues with personal conflicts but it has ESO likely needs to establish a policy on
also been suggested that AI may be more whether LLMs are allowed to be used in

40 The Messenger 194 | 2025

Astronomical News DOI: 10.18727/0722-6691/5386

Report on the ESO workshop

The Promises and Challenges of the ALMA Wideband

Sensitivity Upgrade
held at ESO Headquarters, Garching, Germany, 24–28 June 2024

Carlos De Breuck 1 ALMA user community was informed in involved. While these polls obviously do
María Díaz Trigo 1 detail of the broad scope of the WSU. not pretend to offer a representative
While many of the technical upgrades ­opinion of the whole ALMA user commu-
within the WSU are intrinsically linked to nity, they did help to streamline the dis-
1
ESO each other (for example, digitisers, digital cussion and to provide some guidance
signal transport and correlator), others about community priorities for the WSU
are relatively independent (for example, (see below).
The Atacama Large Millimeter/sub­ receiver upgrades, data reduction soft-
millimeter Array (ALMA) is undergoing ware). Therefore, having scientific input One positive outcome of these interactions
the most ambitious project since its from the community to help set the priori- between technical and scientific groups
inauguration: the Wideband Sensitivity ties within the WSU project was a second was a lively discussion of how ALMA
Upgrade (WSU). The WSU will increase goal of the meeting. users can help increase the visibility of
the instantaneous bandwidth by as instrument builders. For example, encour-
much as a factor of four while retaining The conference brought together the sci- aging astronomers to cite the relevant
full spectral resolution over the entire entific and technical user communities, instrument papers in their scientific papers
bandwidth, thus resulting in increases while also covering a broad range within can help obtain funding for new receiver
of the spectral scan speed of up to a these communities. For example, the development. Facilitating access to instru-
factor of 50 for the highest spectral reso- ­science topics covered the Sun, the Solar ment papers in the SAO Astrophysics
lution. In addition, an upgrade of the full System, protoplanetary discs, stars, Data System, making accessible to users
signal chain of ALMA — from the receiv- astrochemistry, the interstellar medium, a list of relevant technical papers 3 or
ers and digitisers all the way through nearby and distant galaxies, galaxy clus- updating the ALMA citation policy were
to the correlated data — will result in ters, cosmology and special observing discussed in this context.
increases in sensitivity for all observa- modes such as time domain, polarisation
tions. However, the increased bandwidth and very long baseline interferometry.
and throughput bring several technical The technical topics covered the whole Wideband and sensitivity upgrades
challenges. In June 2024, we organised ALMA signal chain from receivers and
the first conference at ESO in Garching their components, digitisers and correlator, The overarching goal of the meeting was
to inform the ALMA community about to the impact of the WSU on ALMA oper- to present in detail the improvements the
all the details of this upgrade. Here we ational aspects such as user support, WSU will bring to ALMA users. To enable
report on the outcome of this meeting. quality control, data processing software the wider bandwidth and increase the
tools and the archive. The presentations digitisation efficiency, ALMA will first need
are available on Zenodo1. a completely new signal chain. The signal
Motivation chain will include new digitisers in all the
ALMA antennas, which will allow any
In its first decade, the Atacama Large Interaction between science and intermediate frequency (IF) in the range
Millimeter/submillimeter Array (ALMA) has technology 2–20 GHz to be covered, providing a
revolutionised our view of the Universe, ­flexible system to serve the upgraded
both near and far (see Díaz Trigo et al., Fostering interaction between science receivers. A new digital signal transport
2024). Serving a very broad range of sci- and technology during a conference is a system will carry the signal to the Opera-
ence topics has been one of the main challenging task, and has been attempted tions Support Facility (OSF), where the
successes of ALMA, leading to a sus- at several previous meetings, when these Advanced Technology ALMA Correlator
tained high demand from the scientific different topics were often clearly sepa- (ATAC) will be installed to offer spectral
community. Being the most powerful rated in the programme. Unfortunately, resolutions of 0.1 km s –1 or better over
(sub)mm observatory in the world also this leads to many participants attending the entire correlated bandwidth, so that
comes with the responsibility to continu- only the session(s) in their own field, which ALMA users no longer have to sacrifice
ously fulfill the high expectations of the does not stimulate interaction between bandwidth for spectral resolution. Even
community. While ALMA set new stand- the groups. We therefore attempted a before upgrading the receivers, the new
ards in (sub)mm technology during con- new structure in the programme2, mixing digitisers and the ATAC will already offer
struction, other observatories such as technical and scientific talks within the an increase of spectral grasp and a sensi-
the Northern Extended Millimetre Array same session. To retain coherence, we tivity improvement of more than 10%.
(NOEMA) and the Submillimeter Array did keep a focused technical and scien-
(SMA) now offer an instantaneous band- tific topic within each session. Each half- When combined with ALMA’s high spatial
width coverage wider than ALMA’s. day was then concluded with a one-hour resolution, the WSU will allow much more
Indeed, expanding that bandwidth was discussion on a topic relevant to both powerful studies of astrochemistry in a
set as the top priority of the ALMA2030 the preceding scientific and technical wide variety of targets including comets,
roadmap (Carpenter et al., 2020), and the talks. To further stimulate engagement of protoplanetary discs, prestellar cores
Wideband Sensitivity Upgrade (WSU) is the participants, we organised quick and (high-mass) protostars, star-forming
the response to that goal. The WSU2024 online polls using Mentimeter, which also regions, and nearby galaxies. Several
conference was the first time the entire allows online participants to be actively speakers highlighted the formidable

The Messenger 194 | 2025 41

Astronomical News De Breuck, C., Díaz Trigo, M., Report on “The Promises and Challenges of the ALMA WSU”

­ hallenge that awaits them in modelling

c commensurate gain in observing time, Figure 1. Conference photo.
WSU datasets containing thousands of but also more uniform data over a wide
molecular lines. Results from several spectral range, thereby increasing their
ALMA Large Programmes presented at archival value. For sources with a lower lenses or cryostat entrance windows
the meeting already gave a first glimpse density of spectral lines, a broader band- may not need upgrades, while others
of what to expect from WSU datasets, width also increases the flexibility in the such as the cryogenic low-noise amplifi-
showing that we are already at the limit of spectral setup. Several speakers high- ers already have prototypes covering IF
what astronomers can handle with the lighted the importance of such more effi- bandwidths from 4 to 20 GHz which
current analysis tools. With the WSU, cient spectral setups, for example for would meet the WSU requirements.
these datasets will increase by at least an time-critical or time-intensive observa- Improving in areas such as the polarisa-
order of magnitude. Significant upgrades tions of variable sources or polarisation tion accuracy may come with a penalty
of the spectral fitting and chemical analysis measurements, for which the use of sub-­ on sensitivity, which could be alleviated
tools are already starting, and will require arrays was also suggested as a way to by reducing waveguide losses using new
the incorporation of machine learning further increase the simultaneous broad fabrication technologies.
techniques. bandwidth. The factor of four increase in
bandwidth will result in an improvement The receiver upgrades presented for
The promise of the WSU is to increase of the sensitivity and calibration accuracy, Bands 6, 7, 8, 9 and 10 are all planning
the instantaneous bandwidth by at least which is important for, for example, accu- to use silicon-insulator-silicon technology,
a factor of two, the goal being a factor rate spectral slope determinations. where the most challenging part is to
of four (i.e., 16 GHz per sideband and per combine a bandwidth four times broader
polarisation). Reaching this goal of a While the ATAC ingest has been designed with tighter requirements on the receiver
­factor of four came out as a clear priority to record up to four times the current temperature and sideband rejection. The
from the quick poll of the participants bandwidth, reaching this full capacity will new Band 2 receiver (extending over the
(see Figure 2). The benefits of the broadest require additional hardware investments. existing Band 3 range) instead uses a
possible bandwidth are obvious for spec- Reaching bandwidths of 16 GHz is also low-noise amplifier as the first active
tral line surveys (both Galactic and extra- a challenge for the receivers being component (see Figure 3). Band 2 is both
galactic, such as redshift surveys), where upgraded. Some receiver components the last of the originally planned ALMA
broader bandwidth provides not only a such as the optics, corrugated horns and receivers, and also the first receiver that

42 The Messenger 194 | 2025

 band1 more bandwidth one. With the diversity of the ALMA user
16 ghz bw wider bandwidths wide frequency range community represented during the
higher spec res at low fr survey speed high sensitivity
WSU2024 conference, it became clear
wide bandwidth continuüm sensitiviy
wide band
invvestment

x4 bandwidth that it is impossible to find a strategy

sensitivity 4x dual frequency
that will satisfy all users. For example,
band2

bandwidth
mapping speed
during the most intense WSU commis-
band9

calibration accuracy
b7 sioning phase, a reduced number of
band 7 4x bandwidth
16 ghz bandwidth antennas and time will be available for

spectral resolution bandwith

speed
more bw Figure 2. Results from
a quick poll of the con-
science observations, and some observ-
ing modes or configurations may not be
4x correlator wider bandwidth technology ference participants to offered. The alternative of a prolonged
quantum noise limit high frequency if bandwith gauge their main priority
phased array mode bw exte spectral resolutiooooooon for the WSU.
shutdown period, or even skipping a full
Cycle, was not considered as attractive
by most of the conference participants.
is compatible with the new WSU require- Astronomy Data Processing System
ments. As the ongoing installation of the (RADPS). The WSU2024 conference also
Band 2 production receivers should be had a session dedicated to collecting Demographics
completed by the time the ATAC, digitisers new ideas and user requirements for the
and digital signal chain are being installed, ALMA archive in the WSU era. It became The Scientific Organising Committee
Band 2 will be a critical band for the WSU clear that there is a strong desire to (SOC) included representatives from all
commissioning. ­execute more of the processing steps the ALMA Regional Centre nodes as
on servers hosting the data, especially well as ESO ALMA staff and the ALMA
During the discussion sessions of the for ALMA users who do not have access observatory scientist. To ensure a rep-
WSU2024 conference, there was con- to powerful computing facilities. In addi- resentation from a wide range of scientific
structive input from ALMA users on the tion, machine learning techniques may and technical fields, the SOC invited
timeline of the upgraded receivers. With help to alleviate the formidable challenges 27 speakers. Both the SOC and invited
the upgrade projects for Bands 6 and 8 of the WSU data volumes. speakers had a 44% female participation.
already started, the participants expressed The total number of registered partici-
a clear preference for Band 7 to be the A major upgrade like the WSU will not pants was 152, of whom 110 attended
next priority, consistent with the priorities come without some sacrifices by the in person and 42 remotely. The overall
originally set in the ALMA2030 roadmap. ALMA users. A mitigation plan to reduce female participation percentage was
Interestingly, the conference participants the operational downtime during the 39%, which could reflect the lower female
appeared to be willing to compromise on upgrade was presented during the meet- fraction in technical staff positions.
the schedule or even forego the last GHz ing. For example, during the commis­ Although the SOC tried to attract more
of bandwidth to see this important band sioning of the initial WSU, the new signal early-career scientists by offering a
upgraded. After band 7, most participants chain will operate in parallel to the existing reduced registration fee, only 28% were
preferred to upgrade band 9 before students or postdocs. While the meeting
bands 1, 4+5 and 10. primarily targeted European users,
30 participants from North America,
East Asia and Chile also participated.
Impact on ALMA users

As previously mentioned, the increased Acknowledgements

spectral grasp and bandwidth of ALMA We thank the Scientific and the Local Organising
in the WSU era will lead to substantially Committee for their extensive and efficient help
increased data volumes. While the in making this conference a success.
median increase in the product size is
expected to be less than an order of References
magnitude, some products will be more
than three orders of magnitude larger. Carpenter, J. et al. 2020, arXiv:2001.11076
This will require the replacement of the Díaz Trigo, M. et al. 2024, The Messenger, 193, 57
current ALMA data reduction software, the
Common Analysis Software Applications Links
(CASA), by a new suite named the Radio
1
Zenodo link to the presentations:
https://zenodo.org/communities/alma-wsu-2024
2
L ink to workshop page:
Figure 3. The ESO-led Band 2 receiver will be the first https://www.eso.org/sci/meetings/2024/wsu.html
3
ALMA receiver compatible with the WSU require- ALMA technical handbook: https://almascience.
ments, covering an IF bandwidth from 2 to 18 GHz. eso.org/proposing/technical-handbook

The Messenger 194 | 2025 43

Astronomical News DOI: 10.18727/0722-6691/5387

Report on the ESO workshop

Galaxies at Crossroads: Outflows and IMF in the

VLT/ELT/ALMA/JWST Era
held at Brno Observatory and Planetarium, Czech Republic, 16–20 September 2024

Tereza Jerabkova 1

Brno Planetarium
Alice Concas 1
Ivanna Langan 1,2
Belen Alcalde Pampliega 1
Glenn van de Ven 3
Jorryt Matthee 4

1
ESO
2
Centre for Astrobiology (CAB),
CSIC-INTA, Madrid, Spain
3
Department of Astrophysics, University
of Vienna, Austria
4
Institute of Science and Technology
Austria (ISTA), Klosterneuburg, Austria

The international workshop Galaxies at

Crossroads gathered a diverse group of
researchers to address pivotal themes
in galaxy evolution, including galactic
outflows, the stellar initial mass function,
and the mass-metallicity relation.
Hosted in collaboration with Masaryk
University, the event featured a robust
scientific programme complemented
by interactive activities, emphasising
collaboration and fostering early-­career
participation. The location in Brno,
close to the University of Vienna and
the Institute of Science and Technology
Austria, provided an ideal and sustain­
able meeting point in central Europe,
easily accessible by train. This report
highlights the science, the participant
statistics, and how effective it can be
when workshops are organised outside
of ESO to foster stronger interactions
between the community and ESO.

Motivations

The workshop was driven by the need to

address pressing questions in galaxy evo-
lution from different angles and to foster
collaboration across different extraga­
lactic research domains. ESO’s strategic
focus on supporting Member States and
strengthening regional scientific commu-
nities was central to its objectives.

A major theme of the meeting was bring-

ing together experts on three intercon-
nected yet expansive topics: galactic
outflows, the stellar initial mass function
(IMF), and chemical evolution. Each of

Figure 1. Conference photo.

44 The Messenger 194 | 2025

these areas is often the focus of separate AGN-driven outflows in dwarf galaxies. prevent fatigue. This ‘light’ programme
conferences, but this workshop sought An invited talk by Jorryt Matthee pre- gave participants time to properly digest
to foster interactions among experts sented future prospects for spectro- the talks and engage in meaningful
across these fields. A particular emphasis scopic observations of galaxies in the ­discussions during the breaks and ded-
was placed on bridging observational first Gyr, followed by group discussions icated discussion sessions. The use
and theoretical perspectives, uniting led by Filipo Fraternali, Antonino Marasco, of different talk lengths — from longer
observers, simulators, and theorists to Ivanna Langan, Bronwyn Reichardt Chu, keynote presentations to concise five-­
encourage cross-disciplinary dialogue. and Matej Barta. minute contributed talks — helped
– Initial Mass Function (IMF): day three strike a balance between giving space
Additionally, the workshop aimed to featured invited talks by Zhiqiang Yan to key topics and allowing more contrib-
enhance engagement in the central on chemical tracers for constraining the utors to present their work. In particular,
Europe region, which hosts rapidly grow- IMF and Alina Boeker on observational the concise talks proved effective for
ing research institutes, and to strengthen and simulation-based IMF studies. conveying key science messages and
ties with ESO and ALMA science. Brno, Highlights included Ankur Upadhyaya’s essential details. This approach was
strategically located between the capital evidence for very massive stars in inspired by the Lorentz Center workshop
cities Prague and Vienna, served as an UV-bright galaxies and Prasad Sawant’s “Gravitational waves: a new ear on the
accessible and sustainable venue, reinforc- unveiling of baryon cycles in z ~ 5 gal- chemistry of galaxies” (29 April – 3 May
ing ESO’s commitment to supporting axies. The discussion session was 2024)2 which demonstrated the value of
research in Member States. organised by Ignacio Martin Navarro, varying talk formats to maintain engage-
Marie Zinnkann, Andrew Hopkins, and ment while maximising impact. Unlike
Glenn van de Ven, fostering interaction programmes packed with numerous
Summaries of talks and highlights from between theorists and observers. talks, which can leave attendees drained
sessions – Mass-Metallicity Relation (MZR): the and skipping sessions, this structure
fourth day began with keynote speaker encouraged full participation. It was
A major theme of the meeting was to set Allison Strom, who re-examined metallic- fantastic to see that everyone stayed
the stage for collaborative discussions ity through chemical abundance patterns engaged throughout, creating a vibrant
across fields that are often tackled inde- in distant galaxies. Additional highlights atmosphere for experts from varied fields
pendently. The programme1 included: included Dirk Scholte’s a ­ nalysis of tran- to interact and collaborate effectively.
sitions in the baryon cycle and Gauri – Discussion Sessions: each scientific
–S  etting the Scene: the workshop opened Kotiwale’s work on galaxy metallicity at topic was paired with a dedicated dis-
with an overview of active galactic the epoch of reionisation. Discussions cussion session. Participants were
nuclei (AGN) and the Czech Space were facilitated by Belen Alcalde, Zhuang divided into groups based on career
­Telescope QUVIK, presented by Michal Zhuyun, Gauri Kotiwale, and Ivanna stage and expertise to ensure balanced
Zajacek, followed by Pavel Jachym’s Langan, focusing on integrating obser- representation, while other discussions
update on ALMA science, Czech ALMA vational and theoretical perspectives. involved all attendees using online tools
Node activities, and jellyfish galaxies. – Final Day Discussions: the workshop for real-time voting and comments.
Participant introductions fostered by concluded with X-ray astrophysics high- – Group Summaries: after group discus-
noughts-and-crosses-style activities lights, including talks by Julia Falcone sions, joint sessions were held, with
(a variation of bingo, which proved to be on Seyfert galaxy feedback and Michal early-career researchers bravely pre-
a great success) during the welcome Zajacek on ultrafast outflows. Norbert senting their group’s findings, fostering
drinks reception created an engaging Werner introduced high-energy astro- inclusivity and confidence.
atmosphere for networking. The goal physics research being conducted at
was to create an environment in which it Masaryk University. This was followed
was easy to engage with new people, by a short wrap-up discussion involving Main conclusions and ways forward
ensuring that participants from three dif- the Scientific Organising Committee
ferent astronomy fields had the opportu- and participants, and a planetarium The workshop achieved its goals of fos-
nity to connect beyond their usual circles. show highlighting synergies between tering collaborations and advancing dis-
– Galactic Outflows: the second day high-energy astrophysics and galaxy cussions on galaxy evolution. Plans to
began with Barbara Mazzilli Ciraulo’s evolution research. build synergies across related fields were
highlight talk on cold outflow filaments emphasised, as this approach is essential
in Generalising Edge-on galaxies and for understanding the full picture, particu-
their Chemical bimodalities, Kinematics Interactive and collaborative activities larly in the context of growing synergies
and Outflows out to Solar environments between facilities, such as the JWST, the
(GECKOS) winds, followed by Alejandro The workshop was designed to ensure Atacama Large Millimeter/submillimeter
Olvera’s exploration of gas flows in active engagement and discussions: Array, and ESO’s upcoming Extremely
NGC 99. Other highlights included Large Telescope (ELT). Strengthened
Bronwyn Reichardt Chu’s discussion – Session Structure: sessions were care- ­networks among early-career research-
of star formation-driven outflows and fully timed to allow a focus on each ers were another significant outcome,
Archana Aravindan’s focus on presentation, with sufficient breaks to with mentorship opportunities fostering

The Messenger 194 | 2025 45

 Astronomical News Jerabkova, T., et al., Report on “Galaxies at Crossroads”

Figure 2. This image is from a sneak preview we

T. Jerabkova

received of the exhibition prepared by the Brno

Planetarium staff, showcasing their exceptional work
and support during the workshop.

– 10% from Asia (including China and

India),
– 5% from South America (notably Chile
and Brazil), and
– 5% from Australia.

The distribution highlights strong engage-

ment from European institutions, with
a growing presence from North America
and Asia, reflecting the international
appeal of the workshop. To further pro-
mote inclusivity, pronoun stickers were
introduced, allowing participants to
­display their preferred pronouns on their
badges. This initiative, which aimed to
create a welcoming environment for all
attendees, has since been adopted by
other ESO workshops, underscoring its
positive impact.

Acknowledgements

The organising committee extends its gratitude to

ESO for funding and workshop support, and to the
ESO Workshop Selection Committee for endorsing
this event. Special thanks go to the Planetarium in
Brno 5 for providing an exceptional venue and out-
standing technical assistance, which were critical
to the workshop’s success (Figure 2). The event also
inclusivity and development. It was par- selected to represent a balance of career benefitted greatly from the contributions of local
ticularly rewarding to witness exceptional stages and scientific expertise. The final students, postdocs, and staff members who pro-
vided invaluable support with the organisation. We
collaboration across fields and between programme included 47% female and acknowledge Masaryk University for its collaborative
early-­career and senior scientists. The non-­binary participation, aligning with the efforts in hosting the workshop. Finally we thank all
group dynamics were notably enriched ­proportion of female and non-binary par- participants for their inspiring attitude and openness
by diverse discussion formats, including ticipants at the workshop. Additionally, to collaborations and we thank all invited speakers.
small group splits and online tools that over 70% of attendees were students
allowed anonymised audience input in and postdoctoral researchers, reflecting Links
real time. Additionally, the event fostered ESO’s emphasis on fostering the next
1
new connections between Czech and generation of astronomers. Early-career  ink to workshop programme: https://www.eso.org/
L
sci/meetings/2024/galcross/programme.html
Austrian institutes, laying the groundwork researchers also contributed as session 2
Lorenz Center Workshop: https://www.
for future collaborations. Recorded talks chairs and discussion leaders, showcas- lorentzcenter.nl/gravitational-waves-a-new-ear-on-
and presentations are available on ing their leadership potential. the-chemistry-of-galaxies.html
YouTube3, and are accessible on Zenodo4. 3
Recorded presentations:
This workshop exemplifies the potential Participants (Figure 1) came from over https://www.youtube.com/playlist?list=PL-
7vLpk0VDrILpaZsWQNntDLVmHav1G7U
for similar collaborative efforts, especially 20 countries, with notable contributions 4
Presentations on Zenodo:
in preparation for the ELT and other next-­ from ESO Member States and partner https://zenodo.org/communities/galcross2024/
generation observatories. institutions. The breakdown of participants records?q=&l=list&p=1&s=10&sort=newest
by region was: 5
B rno Planetarium webpage:
https://www.hvezdarna.cz/en/

Demographics – 65% from Europe (notably Germany,

Italy, Czech Republic, Austria, France,
The Scientific Organising Committee and Spain),
sought fair representation from the – 15% from North America (USA and
­community. Invited speakers were Canada),

46 The Messenger 194 | 2025

Astronomical News DOI: 10.18727/0722-6691/5388

Report on the ESO workshop

New Heights In Planet Formation

held at ESO Headquarters, Garching, Germany, 15–19 July 2024

Anna Miotello 1 Rationale ongoing at different facilities. Theory and

Enrique Macias 1 models are therefore faced with the task
Planet-forming discs can nowadays be of explaining much more complex sce-
probed in unprecedented detail thanks narios of disc evolution, planet formation,
1
ESO to facilities such as the Atacama Large and planet–disc interaction.
Millimeter/submillimeter Array (ALMA) at
submillimetre wavelengths or high-­contrast This workshop aimed to bring together
Planet formation is a widespread by-­ imaging instruments in the near-infrared, observers with expertise in different
product of the process of star formation such as the Spectro-Polarimetric High-­ wavelength regimes, theorists, and mod-
itself and occurs within protostellar discs contrast REsearch (SPHERE) instrument ellers to evaluate our current progress,
made of gas and dust that orbit the new-­ at ESO’s Very large Telescope. In the past pinpoint critical unresolved challenges,
born star. In the past few years new decade these facilities have transformed and discuss ways forward.
observations of discs at various wave- the field of planet formation, enabling
lengths — thanks to new-­generation both moderate-resolution statistical disc
facilities — have revolutionised the field surveys and high-resolution imaging Programme
of planet formation and challenged some studies of discs. Improvements in data
of the traditional theories. This workshop quality and sample size have, however, Several topics were addressed during
brought together around 200 scientists raised many fundamental questions about the ESO workshop. We started with a
from all over the World to review the the structure of discs and their evolution, beautiful and extensive review of recent
state of the art, pinpoint the main open all the way to the formation of planets. JWST results in the context of planet-­
questions, and explore new avenues. This observation-driven field seems to be forming discs, with a special focus on
As well as invited reviews and talks, and continuing along this path with upcoming the level of improvement compared to its
contributed talks, the scientific pro- results from the James Webb Space ­predecessors. This set the stage for a
gramme offered ample space for infor- Telescope (JWST) and the many recently
mal poster viewing sessions. accepted Large Programmes that are Figure 1. Conference photo.

The Messenger 194 | 2025 47

Astronomical News Miotello, A., Macias, E., Report on “New Heights In Planet Formation”

Participants’ career stage Participants’ country

USA France

UK Estonia

Student Taiwan Denmark

Postdoc Switzerland Czech Republic

Staff Spain China

Netherlands Chile

Japan Canada

Italy Belgium

Ireland Austria

Germany Australia

Figure 2. Demographics. contributions, 75 posters were presented, Outlook

divided into two poster sessions in the
series of novel JWST results highlighted first and second halves of the week. The The workshop was the major scientific
in both contributed talks and posters. SOC decided to give more visibility to event in the field of protoplanetary discs
The second day was devoted to the enor- some of the outstanding results presented in 2024. Feedback from the participants
mous contribution of ALMA to our under- on posters by rewarding the three best was very positive. The efforts to give
standing of discs. Some results from two posters with a five-minute talk on the last visibility to ESRs was much appreciated,
of the recently accepted ALMA Large day of the conference. Two of the best as was the friendly atmosphere, which
Programmes (AGE-PRO and Exo-ALMA) posters were selected by a committee allowed interesting and respectful inter-
were presented, addressing the funda- and the third one was voted by the con- actions and discussions. Despite the
mental yet unsolved problem of disc ference participants through a web poll. large number of participants, reaching
evolution and the prospects of detecting the limits of the ESO capacity, the organi-
embedded protoplanets using disc kine- sation was very smooth, mostly thanks
matics. The third day was mostly devoted Demographics by the numerous and engaged Local
to theory and simulations with a special Organising Committee (LOC), composed
focus on planet–disc interaction. On the The workshop was co-funded by the of 12 ESO students, fellows and staff
fourth day we addressed the effects of DUSTBUSTERS1 collaboration on members.
the environment on discs. That topic was protoplanetary discs, funded by an EU
possibly the one that generated most Marie Skłodowska Curie RISE grant. It Slides from most talks, the detailed pro-
interest and stimulated most of the dis- was timed to serve as the closing event gramme, the list of posters, and the LOC/
cussion. Recent results, in fact, have of DUSTBUSTERS. The SOC consisted SOC composition are available at the
shown that external processes may be of DUSTBUSTERS node leaders and workshop website 2.
very relevant in discs even at later stages external scientists, four female and five
while planets are forming. Finally, on the male members, with nine SOC members
last day, the topic of astrochemistry was from seven countries (Australia, Italy, Acknowledgements
addressed and new advances in this area Chile, USA, France, UK, Germany). Our We would like to thank all participants, both in person
thanks to the recently accepted ALMA final numbers included 267 registered and remote, for their active participation in the confer-
Large Programme DECO were presented. participants, of whom 186 were in person, ence, which was crucial to making it such a success.
from more than 20 different countries. We would further like to thank our SOC and LOC
members for their fundamental and invaluable effort.
The programme included 67 talks over A special thanks goes to Denisa Tako for her support
five days, including seven invited review The SOC made every effort to ensure with the organisational aspects of the conference.
talks and five invited talks, the remainder a balanced scientific programme, with
being contributed talks selected by the 27 of 67 speakers (~40%) being female.
Links
Scientific Organising Committee (SOC). Likewise, the priority was to support and
For the five invited talks the SOC selected give visibility to early-stage researchers 1
 ink to the DUSTBUSTERS Webage:
L
junior members of the main five Large (ESRs) in selecting both invited speakers https://dustbusters.fisica.unimi.it/
Programmes ongoing in the field of and contributed talks. In fact, 55 of 67 2
L ink to workshop programme: https://www.eso.
planet formation. We also held one open org/sci/meetings/2024/dustbusters.html
speakers (~82%) were either PhD students
discussion session. As well as the oral or postdocs.

48 The Messenger 194 | 2025

Astronomical News DOI: 10.18727/0722-6691/5389

Report on the ESO workshop

A Decade of Discoveries with MUSE and Beyond

held at ESO Headquarters, Garching, Germany, 18–22 November 2024

Eric Emsellem 1 (MUSE) mounted on Unit Telescope 4 of in their presentations existing and future
Joël Vernet 1 the Very Large Telescope (VLT) quickly synergies with other major facilities such
Pascale Hibon 1 became a reference instrument address- as the Atacama Large Millimeter/sub­
Amelia Fraser-McKelvie 1 ing a rich and wide range of scientific millimeter Array (ALMA) and JWST — and
Ashley Barnes 1 questions. Combined with the powerful soon ESO’s Extremely Large Telescope
adaptive optics facility, MUSE has pro- — as well as to discuss the current
foundly changed the way observers think ­challenges and prospects for science
1
ESO and prepare their observing programmes. supported by integral-field spectroscopy,
It has opened up new avenues into a building on their MUSE experience.
­variety of scientific topics covering, for
The Multi Unit Spectroscopic Explorer example, galaxy formation and evolution,
(MUSE) spectrograph is currently the the nature of the circumgalactic medium, The workshop
most in-demand instrument at the VLT early stellar evolution, and stellar popula-
and since its commissioning in 2014 it tions (see Figures 1 and 2). This ‘MUSE The programme1, designed by the scien-
has served a broad scientific commu- at 10 years’ workshop was organised to tific organising committee (SOC), was
nity covering many research fields in provide a timely opportunity to discuss split into half-day sessions that covered a
astrophysics. As a stable, relatively past achievements, to probe synergies broad set of topics, including the high-z
wide-field, two-dimensional, spectro- between integral-field spectroscopy and Universe, the circumgalactic medium,
photometric facility in the optical, other existing or upcoming facilities, and gravitational lensing, galaxy evolution,
assisted by state-of-the-art adaptive most importantly to address the current planetary systems, the physics of galactic
optics, MUSE has revolutionised our and expected challenges and to nurture nebulae, globular clusters, and stellar
perspective on the use of integral-field potential ideas for the future. populations, as well as the baryon cycle
spectroscopy. This has been accompa- (for example, star formation, stellar-driven
nied by a steep learning curve in the A particular focus of this four-day ESO feedback) and AGN. Several upcoming
community on how to best reduce, workshop was on notoriously difficult and prospective projects were empha-
­analyse and exploit its unique datasets. aspects such as background subtraction, sised, including BlueMUSE and the Multi-­
This dedicated workshop was a unique extraction of spectra in crowded fields, the conjugate-adaptive-optics-Assisted
opportunity to review the scientific low-surface-brightness regime, line spread ­Visible Imager and Spectrograph (MAVIS)
achievements that MUSE has allowed function and point spread function meas- for the VLT and the Wide-Field Spectro-
over the last decade, to better under- urement and homogenisation, astrometry scopic Telescope. Dedicated sessions
stand and reflect on the synergies and mosaicking. Speakers were specifi- were organised around data and tools,
between MUSE and other facilities and cally requested to highlight and address illustrating the MUSE data reduction and
to discuss the associated present and
future challenges it entails. This work-
shop witnessed the gathering of a

Credit: M.-L. Aru/ESO

strong, diverse and interconnected
173–236
community that could report on their 170–249

experience and results and discuss

potential avenues for the future, further
emphasising the benefit of collaborative (NII)Hα(OIII) (NII)Hα(OIII)
developments and shared knowledge.
(NII)Hα(OIII)

θ1 Ori C
Motivations
167–325
177–341W
168–326
With its large field of view, broad wave- 168–328

length coverage, state-of-the-art adaptive 244–440

optics, and spectrophotometric capabili- (OI)Hα(SII) 154–346

ties, the Multi Unit Spectroscopic Explorer
H
(OI)Hα(SII) (ArIII)Hα(OIII)
174–414

170–334
Figure 1. Insets: Reconstructed RGB images of con-
tinuum-subtracted, single-line integrated flux images 170–337
of the sample of proplyds in a star formation region θ2 Ori A
acquired with MUSE (Aru et al., 2024; Haworth et al., 203–504
169–338
2023). In each inset, various emission lines are com-
bined to highlight the morphology of the proplyd. 171–340
(NII)Hα(OIII)
Background: Colour-composite using fluxes of three (NII)Hα(OIII)
emission lines, with blue: Hβ, green: (N II) 6584, and
red: (S II) 6731 (Weilbacher et al., 2015).

The Messenger 194 | 2025 49

 Astronomical News Emsellem, E., et al., Report on “A decade of Discoveries with MUSE and Beyond”

Figure 2. Top panel: Image of ω Cen. A three-colour

–47°24' RGB image of ω Cen created from MUSE WFM
data using synthetic SDSS i, r, and g filters (Nitschai
et al., 2023). The image displays the coverage of
all WFM data (both GTO and GO). The red cross
­indicates the centre of the cluster. Bottom panel:
Chromosome/pseudo-colour diagram for a sample
of 7277 RGB stars in ω Cen, each star coloured by
26' its metallicity. RGB stars are separated into three
distinct streams using diagonal black lines and colour-­
coded labels. The edges of the [Fe/H] bins are indi-
cated by white lines on the colour bar. Spreads in
the ∆_{F275W,F814W} and ∆C_{F275W,F336W,
F435W} within metallicity bins are primarily due to
Declination (degrees)

light element abundance variations (see, for example,

Clontz et al., 2024).
28'

The success story

One of the many highlights of this confer-

30' ence was the keynote speech by the PI
of the MUSE instrument, Roland Bacon,
reporting on the “Anatomy of a Success”
and the lessons learned from both the
design and construction phases and the
decade of operational life of the instrument.
32'
The productivity and popularity of the
MUSE spectrograph were clearly empha-
sised via various metrics, with a lower
limit for the pressure between five and
201° 45' 40' 35' 10 depending on the allocation period.
Right ascension (degrees) Some of the reasons for this success were
analysed in the context of, for example,
[Fe/H] its broad scientific capabilities (see also
< –2.25 –2.0 –1.75 –1.5 –1.25 > –1.0 Roth, 2024), performance and discovery
power. As mentioned by Roland Bacon
0.50 in his talk, MUSE and the ground-layer
adaptive optics (GLAO) opened a window
for very long integrations on the sky with
∆C F275W,F336W.F435W

0.25
excellent final image quality for an optical
ground-based facility. One key aspect that
0.00 Uppe r allowed the realisation of such opportuni-
stream
ties is associated with the vision attached
Mid str to instrumental and software develop-
eam
–0.25
L owe r ment, promoting new approaches to the
stream
N* = 7277 design, construction and operation of a
–0.50 spectrograph. MUSE is the result of the
–0.5 0.0 0.5 1.0 1.5 2.0 efforts of a team of hard-working and
∆ F275W,F814W creative people who had a science-driven
vision for an ESO VLT instrument.

analysis and how it connects to the ESO in-person attendees (Figure 3). The SOC It is worth noting here that MUSE’s capa-
ecosystem. In a special 90-minute session, managed to achieve a gender balance bility to blindly target a field, as applied,
Peter Weilbacher introduced the ins and among the invited reviewing speakers, for example, to the MUSE Deep Fields
outs of MUSE data reduction, and Amelia with just over a quarter of junior scien- (the talks by Bacon, Fumagalli, Wisotzki,
Fraser-McKelvie led a discussion on an tists. A blind (anonymised) selection of Ciocan), was a game-changer but did
organised ‘Data Challenge’ (see below). talks led to a majority (60%) of junior not initially receive strong support from
speakers, with a 30:70% ratio of female early reviewers.
The conference gathered more than 140 to male presenters (representative of
scientists and engineers, with 110 local the application pool). Today, the most requested MUSE mode
is the Wide-Field NO-AO mode, which

50 The Messenger 194 | 2025

has also been the most operated one at The MUSE Data Challenge Figure 3. Conference picture of the participants in
front of the ESO Supernova Planetarium & Visitor
the telescope except for period P114. The
Centre, Garching.
relative under-use of the AO mode trig- While a remarkably solid reduction pipe-
gered discussions throughout the week, line has been developed and maintained
with the idea that the superb performance by the MUSE Consortium and ESO to
of MUSE with the GLAO and its ease of address the complex (multi-CCD) MUSE and relevant to the community over the
use may need to be further promoted. dataset, issues remain that impact coming decades.
observing campaigns to various degrees.
This workshop also highlighted how the Some of these are common challenges The organising team, led by Amelia
collaboration strategy from the MUSE faced by most/all integral-field spectro- Fraser-McKelvie, provided the data chal-
GTO team (such as the organisation of graphs and have been addressed in vari- lenge entrants with a set of raw object
Busy Weeks) has proven successful and ous ways by different teams. Chief and calibration frames for two targets.
is a great example for future instrument among these, for MUSE at least, are sky Participants were encouraged to use any
consortia. subtraction and flat-fielding. These prob- tools and techniques at their disposal
lems will only become more prominent (along with freedom of access to the entire
A fair fraction of the 66 astrophysics talks as the community continues to push ESO archive) to respond to two challenges:
demonstrated the powerful synergies MUSE to its limits over the next decade, 1) to remove flat-fielding signatures from
between MUSE and other observational specifically via its allocated Large Pro- resultant reduced data cubes, and 2) to
facilities such as ALMA, the Hubble grammes: for example, the extraordinarily illustrate their best sky subtraction strate-
Space Telescope, JWST and Chandra. faint surface brightness limit imposed by gies. Despite the very tight schedule
This concerns, for example, deep and the Generalising Edge-on galaxies and and the effort required, three scientists
cluster surveys that have been systemati- their Chemical bimodalities, Kinematics belonging to different teams responded
cally covered by multi-wavelength cam- and Outflows out to Solar environments to this call (Tania Urrutia, Jesse van de
paigns, allowing the probing of a large set (GECKOS) programme, or the large Sande and Johan Richard), and the
of galaxies and structures or targeted ­contiguous sky regions that need to be results of their work were collated by the
observations where MUSE serves as a covered with exceptional flux calibration SOC/LOC and presented to the whole
spectroscopic probe via multiple line and that begins with sky subtraction by, for audience during the Data Challenge
continuum tracers. We also witnessed example, the Physics at High Angular ­session. Each entrant provided creative
a few more technical or method-oriented resolution in Nearby GalaxieS (PHANGS) and effective solutions to the set chal-
talks addressing, for example, the deriva- survey and MUSE and ALMA Unveiling lenges, a summary of which can be found
tion of the point spread function pro- the Virgo Environment (MAUVE). The SOC in the slides for this session (to be availa-
duced during AO observations, and the and the local organising committee (LOC) ble on Zenodo). A discussion of the bene-
use of machine learning algorithms to of the MUSE 2024 conference launched fits and drawbacks of each approach,
classify sources or push the extraction of a ‘MUSE Data Challenge’, focusing on comparing various techniques in an ‘apples
information from MUSE datacubes. It those two most pressing issues that will to apples’ manner with a standardised
would be impossible to cover all of the be key to further expand on the innovative set of data followed. This further triggered
fine scientific results presented at the science MUSE has so far delivered. Those the sharing of ideas to better address
conference here. Still, we encourage issues apply to most modern integral-field those and similar challenges, and further
readers to explore the presentations that spectrograph instruments and future facili- led to action items within the community
will be made available via Zenodo. ties, so solving these problems is timely and at ESO.

The Messenger 194 | 2025 51

 Astronomical News Emsellem, E., et al., Report on “A decade of Discoveries with MUSE and Beyond”

Looking forward, together with the to exploit this potential via the development References
community of ideas and tools, to ultimately push the
Aru, M.-L. et al. 2024, A&A, 687, A93
instrument’s limits even further. The ESO Clontz, C. et al. 2024, ApJ, 977, 14
The ESO MUSE 2024 conference provided community and ESO itself should nurture Haworth, T. J. et al. 2023, MNRAS, 525, 4129
a glimpse of the strong scientific commu- these aspects to extend the synergetic Nitschai, M. S. et al. 2023, ApJ, 958, 8
Roth, M. 2024, RNAAS, 8, 54
nity behind this amazing facility. It was potential of multi-wavelength and multi-­
Weilbacher, P. M. et al. 2015, A&A, 582, A114
pervaded by the unique feeling of a facility science and to fully prepare for the
strong identity associated with the MUSE arrival of the next generation facilities.
instrument and science, ensuring a large Links
audience throughout the week despite 1
Workshop programme: https://www.eso.org/sci/
the wide range of scientific themes that Acknowledgements
meetings/2024/muse24.html
were covered (from planets to cosmology). We sincerely thank all the members of the SOC and
The conference also highlighted the spirit LOC and the ESO IT and logistics staff for their hard
of collaboration around MUSE data and work and support. We would also like to thank
science (exemplified, for example, during Michael Hilker and the Garching Office for Science
for their unconditional support. Particular thanks go
the Data Challenge session), with a desire to Denisa Tako for her patient and positive contribu-
tion to all aspects of the organisation.
Roland Bacon/ESO

Inside the UT4 of the Very Large Telescope, part of sharpness and dynamic range of images using the
the Adaptive Optics Facility (AOF), the four Laser AOF-equipped MUSE instrument will dramatically
Guide Stars Facility, points to the skies during the improve future observations.
first observations using the MUSE instrument. The

52 The Messenger 194 | 2025

Astronomical News DOI: 10.18727/0722-6691/5390

Fellows at ESO

Marta De Simone

"[…] I grew up in a cruel land where snow

mixes with honey
And good people wear crowns of thorns
on their heads
Since childhood, I have learned the differ-
ence between blood and wine,
that a life can be broken for a piece of
meat or bread
And I can’t get used to all this happiness
[…]”
Brunori Sas

These words resonate with me deeply, as

I grew up in a small town in Calabria
(southern Italy), where the snow shines in
winter, and the sky is so dark you can
see the Milky Way in all its colours. I was
raised in a modest family that valued
­culture and study, encouraging me to
explore my curiosity and seek knowledge.

One evening, when I was about 12,

I noticed a bright point of light in the sky,
just above a mountain on the horizon.
It was not a streetlight or a plane and
I couldn’t stop thinking about it. I was
curious and seeking an explanation and
my mother, who did not know what it
was, encouraged me to investigate.
Through my research, I slowly realised
I was looking at Venus, and the discovery
blew me away. Beyond that, I was always
mesmerised by the night sky. On drives
home from the mountains, I’d press my
face against the car window, capturing
every second of that amazing starry sky.
Since then, I started learning the constel-
lations and spending every chance I could
looking up.

At the same time, I was always drawn to

music. I started playing the flute, following
in my sister’s footsteps, and dreamed of
attending the conservatory. But my mother no matter where life took me, I always pursuing physics felt intimidating. I had
encouraged me to think carefully, remind- looked for one to join. Singing became my struggled with it in high school, changing
ing me of the challenges my sister had escape valve, my way to release emotions. teachers every year, none of whom were
faced. On the other hand, my sister tried particularly inspiring. However, in my
to push me to carve my own path. In As high school came to an end, I had final year, a new physics teacher arrived
hindsight, I thank them both, because I to decide what came next. In southern showing us physics in a new light and
later realised that my true passion in music Italy, university was the natural step after speaking of vast research frontiers. When
was not playing an instrument but sing- high school, as the job market offered I shared my interest in astrophysics, he
ing. That became clear when, almost by few alternatives. But I still had no clear and my maths teacher encouraged me to
chance, I joined the church choir in my direction. I considered scientific criminol- go for it. Their support gave me the cour-
home town. At first, I was extremely shy ogy (to solve homicides and mysteries), age to try. I moved to Florence, 800 km
and promised myself I would only ever sing medicine (inspired by my father, who was from home, to pursue a bachelor’s in
within my room’s walls. But that promise a nurse), and, finally, something related physics. The journey wasn’t easy, but worth
did not last long. I never left the choir, and to stars (as astrophysics). But the idea of it. Over time, even my male classmates,

The Messenger 194 | 2025 53

Astronomical News De Simone, M., Drevon, J., Artur de la Villarmois, E., Fellows at ESO

skeptical and envious, came to recognise I then pushed them to create a small
my determination and the difficult path I choir to have a place where people could
had chosen, one they hesitated to follow. sing, relax, and unwind after long days
of work. Two years later, the choir is still
At university, I didn’t even fully under- going, and it is very well-received and
stand what a doctorate was, nor the path full of the enthusiasm of the participants.
to an academic career. That changed
when I had the incredible opportunity, Today, I feel privileged to be an astrono-
after meeting with Leonardo Testi at the mer at ESO, and I hope to always retain
University of Florence, to carry out my the astonishment and curiosity I felt at the
bachelor thesis project at ESO. It was my beginning of my research journey. Now,
first approach to the world of research, I aspire to be like those who inspired me
and the opportunity to see what an inter- at ESO, and to guide and encourage
national institute running the most power- the future generation of astronomers.
ful ground-based telescope looked like.
I fell in love with the place, the inspiring
people working there, and the atmos- Julien Drevon
phere of curiosity. This sparked a desire
to return for my master’s thesis. Soon My name is Julien Drevon, I am 27 and
after, I was offered a PhD position on I grew up in a loving home in the city of
astrochemistry, the topic of my bachelor Cannes, France, right in the heart of the
thesis, to work with one of the leading French Riviera. The sun, the sea, the
experts in the field, Cecilia Ceccarelli. mountains, and the clear skies provided
I was overjoyed and truly grateful for this by the nearby Alps created the ideal con-
opportunity. During my PhD, I studied ditions for nurturing a passion for astron-
how molecules form in space and how omy and the many celestial objects of
the building blocks of life might have the night (and day) sky.
reached Earth, observing the gas in
young stellar nurseries. It was a dynamic Like most people working in the field, my
and enriching journey filled with people passion began at a very young age. If
from diverse backgrounds, wide range you asked my mother, she would tell you
of topics with unexpected and fascinating that I hadn’t even mastered the French
ramifications and changes of directions. language yet, but I was already trying to environments of evolved stars in the infra-
It allowed me to grow into an independ- express my fascination with the Moon red using interferometric observations.
ent scientist, ready to compete for a every time I saw it.
prestigious ESO fellowship. During the second year of my PhD, I had
After a standard academic path and upon the pleasure of being selected for the ESO
Joining ESO as a fellow was a dream turning 17, I followed the usual trajectory Studentship programme in Santiago. At
come true. It brought me back to the of French students pursuing higher edu- ESO, I was mentored by experienced
place that had captivated me and made cation. I completed a year of preparatory astronomers who shared their knowledge
me fall in love with research. As a fellow, studies for the competitive entrance and passion, allowing me to conduct
I spend most of my time studying the exams to the prestigious Grandes Écoles, research and gain expertise that I would
origins of chemical diversity in regions which I passed. This experience taught not have been able to acquire without
where stars and planets form. I could me valuable study methods, but the pace their guidance.
also support the observatory, contribut- and teaching style did not suit me. So,
ing up to 25% of my time with functional I decided to continue my second year After completing my PhD, I was fortunate
work at the ALMA Regional Center, gain- directly in a physics bachelor’s programme to be selected by the ESO Fellowship
ing invaluable insights into how an inter- at the university. The unlimited access to committee to become an integral part of
national major observatory runs. One of knowledge, group work, and independent the team of astronomers and researchers
the most unforgettable moments was the learning were exactly what I needed. based in Santiago. For over a year now,
visit the ALMA site in Chile, where I con- From that point on, my academic journey I have had the privilege of wearing two
ducted observing runs and witnessed followed a more traditional route. hats: I am both a researcher and a night
the magnificence of the 60 antennas on astronomer.
the Chajnantor plateau at 5000 metres. I earned my bachelor’s degree in physics
I once again felt like that child with my face with an astrophysics specialisation, fol- As an ESO researcher, I strive to under-
pressed against the car window, soaking lowed by a master’s degree and a PhD stand how and under what conditions
in every moment of that dark, but colourful in astrophysics at Université Côte d’Azur dust forms around dying stars. Dust and
starry sky. At ESO I also met a few stu- in Nice. My area of expertise focuses the heavy elements produced by stars
dents who shared my passion for singing. on studying the molecular and dusty are the fundamental building blocks of life

54 The Messenger 194 | 2025

as we know it on Earth. By understanding an ESO committee. Once a proposal hometown, to study astronomy. I think
how this dust forms, we are essentially is approved and the target is observable, Argentina is one of the few countries that
trying to understand our own origins. To my role is to provide the best possible data has an astronomy course that lasts five
observe the objects and environments from one of the most optimal observing years, unlike other countries where one
I study, I need telescopes over 100 meters sites on Earth. studies physics and then specialises in
in diameter. This is where interferometry astronomy. I really enjoyed my student life
comes into play: by combining multiple I see this dual role as a significant advan- and, with some ups and downs, I man-
telescopes, we can achieve a resolution tage, especially as someone who com- aged to finish my course, my thesis being
equivalent to that of a single 100-metre pleted their PhD just over a year ago. On about infrared jets in forming stars.
telescope. At Cerro Paranal, we do this the one hand, I have the opportunity to
using either the 8-metre telescopes or the experience full research autonomy, collab- After graduating, the next step was to
smaller 1.8-metre telescopes. orate internationally, and work on topics apply for a PhD position with funds from
I am passionate about. On the other CONICET, the national scientific and
Large diameters are useful when observ- hand, I gain hands-on expertise in inter- technical research council. But, after apply-
ing faint objects that require collecting ferometric data observation, learn the ing for two years in a row, I didn’t win the
large amounts of light to detect their sig- inner workings of an observatory, and grant. That was a difficult moment and
nal. However, the objects I study are develop my skills as a night astronomer. I had to sit down, reflect, and explore
bright enough to be observed with the other paths. I began teaching maths and
1.8-metre telescopes. An additional This is an opportunity unlike any other, physics to first-year university students
advantage of these smaller telescopes is and I feel incredibly fortunate and proud while applying for a master’s programme
their mobility; they can be repositioned to share this experience, thanks to ESO. in Rio de Janeiro, Brazil. I won that grant
at different locations on the platform, and, for the first time, I moved to a differ-
allowing us to scan an object from multiple ent country — and one with a different
angles and achieve resolutions ranging Elizabeth Artur de la Villarmois language.
from a few metres to the equivalent of
a 200-metre telescope — five times the My love for the Universe started in my During my two years in Brazil, I began
resolution of ESO’s Extremely Large childhood, when watching documentaries working with submillimetre observations.
Telescope (ELT) currently under construc- about the origin of the Universe, the Solar My supervisor at that time put me in
tion at Cerro Armazones. System, and Earth’s formation, filled with ­contact with an expert in the field, and I
pretty images and captivating simulations. feel like that was a pivotal moment in my
Interferometry does not directly produce At some point my parents bought me a scientific career. At that time, ALMA was
images of our environment. Instead, we set of little stars that shine during the in its cycle-2, already making revolutionary
observe the intensity and phase of the night, and I glued them on the ceiling of
interference patterns between the light my room, so every night I fell asleep
collected by different telescopes. My ‘under the stars’.
expertise involves gathering, calibrating,
and analysing these data. If enough obser- I was born in San Pedro de Jujuy, a city
vations are available, I can reconstruct located in the north of Argentina, and
an image of the environment surrounding I spent most of my childhood in Salta,
an evolved star. two cities very close to each other and
at a similar latitude as ALMA (just on the
In July 2024 I won, on behalf of ESO, the other side of the Andes). Before finishing
image reconstruction contest at the SPIE high school, I really wanted to study
2024 workshop. This competition, along astrophysics. This wasn’t something tem-
with the scientific challenges presented porary that faded with time and, as uni-
by my data, continuously pushes me to versity application time approached,
refine my skills and improve my image I had to get serious and started consider-
reconstruction techniques. My long-term ing my possibilities. Coming from non-­
goal is to help the interferometry commu- scientific family, I didn’t know what were
nity reconstruct images more efficiently the steps to follow to become a scientist
by making the process less tedious and or what kind of future to expect. I just
more accessible so that more researchers knew I wanted to be an astrophysicist.
can obtain high-quality images. I still remember that my parents were a
little worried about my future, but they
My second role at ESO, given my exper- still supported me and I will always be
tise, is to operate the VLTI instruments grateful for that.
at Paranal for 80 nights a year. I conduct
observations for astronomers worldwide My adventure began when I moved to
whose proposals have been selected by the city of Córdoba, 900 km from my

The Messenger 194 | 2025 55

 Astronomical News De Simone, M., Drevon, J., Artur de la Villarmois, E., Fellows at ESO

discoveries. A few months before finishing My plan was to start my postdoc in April Argentina is not an ESO Member State,
my master’s, I saw a PhD position adver- 2020, but COVID-19 hit just before. I had so it was a lovely surprise when I was
tised in Copenhagen, Denmark, in my started my visa process to come to Chile, accepted at ESO.
research area and fully focused on ALMA but the Chilean embassy in Copenhagen
data. The professor offering the position stopped all the processes until August, I began my ESO fellowship in May 2023,
was the author of several papers that when flights partially resumed. I managed and I feel I have learned a lot in this period
guided me during my master’s so, even to arrive to Chile in August 2020, in the of time, especially about the observational
though I had no idea at that moment middle of the lockdown. The first months aspects of ALMA and how things happen
about Denmark, I really wanted to work were challenging, mainly because of behind the scenes. Visiting the antennas
with him. Fortunately, he offered me the paperwork and bureaucracy, but I was at 5050 metres for the first time was a
PhD position! closer to my home country, so the culture dream come true, and I still enjoy it a lot
shock was not a big deal. every time I have a shift at the high site.
Living in Denmark was a huge change for But not everything is ALMA; in late 2023
me, in terms of culture, distance, people, Living in Santiago I fell in love with the I visited Paranal and the ELT, which was
weather, politics, food, and friends. I fell mountains and with the man who later halfway through construction. That was
in love with the bicycle culture, which I became my husband. During these three breathtaking!
still enjoy today. During my three years of years, I learned more about ESO and
PhD, I began submitting ALMA proposals, ALMA by attending workshops and con- My current research focuses on under-
I wrote first-author papers, I became ferences in Vitacura. I remember the first standing how stars and planets, like our
more skilled in submilletre/millimetre data time I walked outside the Vitacura campus, own Solar System, form and evolve. Using
analysis, and I felt passionate about astro- where there are many beautiful images ALMA observations, I study the chemical
chemistry. After my PhD, I applied for of the Universe and the telescopes, and composition of these systems and their
postdoctoral positions in many countries I thought “I would love to work in this potential to harbour life in the future. Many
like Sweden, Germany, Chile, and the USA. place.” So when my FONDECYT postdoc unanswered questions remain, but the
I was awarded a three-year grant in Chile, was near its end, I applied for an ESO future looks bright, with upcoming new
as a FONDECYT postdoc at Pontificia ­fellowship to work with ALMA. Honestly, capabilities like the ALMA 2030 upgrade
Catholic University (PUC), in Santiago. I thought my chances were low, given that and the ELT’s first light. Stay tuned!
ESO/G. Vecchia

Bathed in the sunset light of Chile’s Atacama Desert,

this image shows steady progress in the construction
of ESO’s Extremely Large Telescope (ELT).

56 The Messenger 194 | 2025