1. Astronomy in Brazil
Science impact of HAWK-I
Mid-infrared imaging of evolved stars
The Carina dwarf spheroidal galaxy
The Messenger
No. 144 – June 2011

2. The Organisation
Brazil’s Route to ESO Membership
Albert Bruch1
1
Laboratório Nacional de Astrofísica –
LNA, Itajubá, Brazil
On 29 December 2010, in a ceremony
held at the Ministry of Science and
Technology in Brazil´s capital, Brasília,
the then Minister, Sergio Machado
Rezende and the ESO Director General
Tim de Zeeuw signed the accession
agreement by which, pending ratifica-
tion by the Brazilian Congress, Brazil
becomes the 15th ESO Member State
and the first non-European member.
An overview of the historical back-
ground, the current state of astronomy
in Brazil, and the motivation that made
Brazil apply to become an ESO Member
State is presented.
Figure 1. A contemporary engraving by Zacharias 1919 in Sobral, Ceará, which contributed
Wagener of a building in 17th century Recife that
History decisively to the first observational proof
may have hosted the first astronomical observatory
in the Americas. for Einstein´s theory of general relativity.
The signature of the accession agree­
ment to ESO (see de Zeeuw, 2011) is the Astronomy began to establish itself in
latest highlight in Brazilian astronomy’s Janeiro (Videira, 2007), and is shown in other Brazilian institutions in the late
very long and distinguished history, which Figure 2. It was originally meant to 19th century and this progressed, primar­
goes back much further than most non- provide essential services to the newly ily in the universities, and most notably
Brazilian astronomers are aware. Long founded state such as time-keeping, in São Paulo and in Porto Alegre, at a
before Brazil was established as a state, and fundamental scientific research in rather modest pace during much of the
at a time when various European pow- astronomy only gradually became part 20th century. Astronomy in Brazil has
ers still disputed dominion over its vast of its activities. Arguably the observa- only really taken off during the past three
expanses, Brazil hosted the first astro­ tory’s most notable scientific achieve­ or four decades. The three main factors
nomical observatory, not just in the ment was the organisation of the expe- that have contributed to this substantial
Americas, but also in the southern hemi­ dition to observe the solar eclipse of and very successful increase are:
sphere. In 1639 the German naturalist and
astronomer Georg Marcgrave founded
an observatory in Recife (Prazeres, 2004), Figure 2. The early
MAST
home of the Observa­
which was then the capital of a Dutch
tory Nacional in Rio de
colony. The probable appearance of this Janeiro, photographed
observatory is shown in Figure 1. Should in 1921, which today
we consider this as the first “European hosts the Museum of
Astronomy and Related
Southern Observatory”?
Sciences.
However, troubled times and warfare
between the countries disputing he-
gemony over the rich Brazilian colonies
impeded the long-term survival of these
initial astronomical activities and astron­
omy only took firm root in Brazilian soil
after the country became an independent
empire in 1822. On 15 October 1827,
the Emperor Dom Pedro I established the
institution that has now evolved into the
Observatório Nacional (ON) in Rio de
2 The Messenger 144 – June 2011

3. 1. New funding lines that permitted
LNA
promising Brazilian students to receive
their professional education abroad,
mainly in Europe and in the USA.
2 Newly created graduate courses in
astronomy, meaning that scientists
could be trained in Brazil, taking
advantage of the expertise brought
back by others who had obtained
their degrees in foreign countries.
3. The creation of the Observatório do
Pico dos Dias (OPD) and the installation
of a medium-sized (at the time) tele­
scope which gave the growing astro­
nomical community access to a com­
petitive observational infrastructure for
the first time.
The above-mentioned factors resulted
in the dramatic growth of Brazilian
astronomy, both in terms of the number
of scientists, as well as in scientific out­
put. It quickly became evident that the Figure 3. Aerial view of the OPD, the principal obser­ purchase has proved to be a severe limi­
vatory on Brazilian territory, located in the Serra da
available instruments were insufficient to tation. Brazil currently owns 2.5% of
Mantiqueira in the southern part of the State of Minas
satisfy the rapidly growing demand and Gerais, operating a 1.6-metre telescope (main build­ Gemini. In 2010 it purchased additional
that there is no really good site for a ing) and two 0.6-metre telescopes. observing time from the United Kingdom,
modern optical observatory in Brazil. So, increasing its access to the telescopes
instead of enlarging the existing facili- angle formed by the cities of São Paulo, by a factor of two and it is anticipated that
ties at a location that is far from ideal for Rio de Janeiro and Belo Horizonte was Brazil will increase its share in Gemini to
astronomical observations, and following chosen as a compromise between easy about 6 % after 2012, when the UK leaves
the modern trend towards the globali­ accessibility and good observing con­ the partnership.
sation of science, it was recognised that ditions. The observatory is operated by
international collaborations were the the National Astrophysical Laboratory With abundant access to rather small
right way forward for the further develop­ (Laboratório Nacional de Astrofísica telescopes (OPD) and limited access to
ment of astronomy in Brazil. [LNA]), based in Itajubá, Minas Gerais, big telescopes (Gemini), the Brazilian
which is a research institute of the Minis­ astronomical community felt the need for
try of Science and Technology, and is something in between: a decent amount
Infrastructure for astronomical research responsible for providing the optical of time at an intermediate-sized tele­
astronomical infrastructure to the entire scope. So Brazil joined forces with three
So as to make telescopes and instru­ scientific community. Today the OPD US institutions (NOAO, University of North
ments with a wide range of apertures, hosts three telescopes with apertures Carolina and Michigan State University)
characteristics and capabilities avail- between 1.6 metres and 0.6 metres, and to build and operate the SOAR Telescope
able to Brazilian astronomers, Brazil it is equipped with an instrument suite (Southern Astrophysical Research Tele­
became a partner in the Gemini Obser- that is tailored to serve its users well. An scope), located next to Gemini South on
vatory, the SOAR Telescope and finally effort to upgrade the observatory is Cerro Pachón (see Figure 4). SOAR is
entered into a Cooperative Agreement underway to keep it competitive, despite a 4.1-metre telescope that is optimised
with the Canada–France–Hawaii Tele­ the increasing light pollution and the for high image quality. Brazil entered
scope (CFHT), giving Brazilian astrono­ growing number of other facilities that are this consortium as the majority share­
mers access to a range of facilities now open to Brazilian astronomers. holder with a stake of about 34%. Brazil­
besides the Brazilian OPD. ian astronomers also have access to
The Gemini Observatory operates two the 4-metre Blanco Telescope at CTIO
Brazilian observational optical astronomy 8-metre-class telescopes on Hawaii through an agreement with NOAO
takes place primarily at the OPD (shown (Mauna Kea) and in Chile (Cerro Pachón) about the exchange of observing time,
in Figure 3). When the observatory was on behalf of a consortium of seven which complements the services and
planned in the 1970s, logistical consider­ countries. Although very well used by instruments offered by SOAR.
ations demanded that the observatory Brazilian astronomers, and extremely
was built within easy reach of the big important for the development of optical The cooperative agreement with the
population centres where most astrono­ astronomy in Brazil, the rather small CFHT, which is located on Mauna Kea,
mers were located. A site within the tri- share of Gemini that Brazil was able to Hawaii, is meant to provide access to a
The Messenger 144 – June 2011 3

4. The Organisation Bruch A. et al., Brazil’s Route to ESO Membership
Figure 4. The 4.1-metre highly productive wide-field 4-metre-class
SOAR Inc.
SOAR Telescope on
telescope with competitive instruments in
Cerro Pachón, Chile.
the northern hemisphere. The agreement
is limited in time and will be reviewed,
with the aim of potentially renewing the
contract, in 2012.
Brazilian participation in all these interna­
tional observatories is managed by the
LNA, which thus exercises a key role in
optical astronomy in Brazil. Apart from
these installations, which are open to the
entire astronomical community, some
institutions operate their own facilities on
a more modest scale, and these either
serve a specific scientific purpose or con­
centrate on education and outreach.
The most recent and arguably most im­
portant (and certainly the biggest) of
these is IMPACTON, a robotic one-metre
telescope for observations of near-
Earth objects, which is currently being
commissioned by the Observatório
Nacional, and is located in the interior of
Pernambuco State.
Other areas of astronomical research
have also benefitted from Brazil’s
contributions to international projects
and collaborations. These include
space astronomy (Brazil is a partner in
the CoRoT space mission, and it is also
engaged in the PLATO mission), high
energy astrophysics (through the partici­
pation of Brazil in the Auger experi-
ment), and cosmology (Brazilian institu­
tions are members of the International
Center for Relativistic Astrophysics Net­
work [ICRA-Net]).
José-Williams Vilas-Boas
The growing importance of large sur-
veys and the exploitation of data banks
for astronomical research has been
recognised and has led to the recent cre­
ation of the Brazilian Virtual Observatory
(BraVO), as the national branch of the
International Virtual Observatory Alliance.
BraVO unites researchers from various
institutions in a coordinated effort to cre­
ate infrastructure and tools for data-
mining and to disseminate the concept
of the Virtual Observatory in Brazil. In
parallel, the LIneA (Laboratório Interinsti­
tutional de e-Astronomia) collaboration
is formed by scientists working at three
research institutes of the Ministry of Sci­
ence and Technology (MCT) to develop
Figure 5. The dome of the Itapeninga Radio Obser­ the infrastructure and software to store
vatory (ROI) in Atibaia, São Paulo.
and process large astronomical datasets.
4 The Messenger 144 – June 2011

5. Figure 6. Mounting the for Gemini, where it was responsible
LNA
1300 optical fibres of
for the fibre feed between the telescope
the SOAR Integral Field
Spectrograph at the and the bench spectrograph (although,
LNA Optics Laboratory. unfortunately, through lack of funding the
instrument was never built). In a success­
ful attempt to find a place on the interna­
tional market for astronomical instrumen­
tation, the LNA has also built the fibre
feed for the Frodospec spectrograph at
the Liverpool Telescope on La Palma.
Independent efforts in instrument devel­
opment are ongoing at the Observatório
Nacional, which, in collaboration with
the IAG, is building a camera for the
J-PAS (Javalambre Physics of the accel­
erating Universe Astrophysical Survey)
project in Spain. Facilities for instrumen­
tation development are also being
Radio astronomy, which was already the Earth from space), a group at INPE installed at the Federal University of Rio
comparatively well developed before the is currently building MIRA X, a small Grande do Norte in Natal.
steep increase in optical astronomy ac- survey satellite to observe the spectral
tivities began, has not followed the same and temporal behaviour of a large num­
steeply rising path. Apart from some ber of transient X-ray sources. Moreover, Size of the Brazilian astronomical
modest investments in specialised instru­ INPE is collaborating with the LNA and community
ments operated by small groups, no the Instituto de Astronomia, Geofísica e
major effort has been made to provide Ciências Atmosféricas (IAG) of the Uni­ According to a census (updated in 2010),
access to a competitive infrastructure for versity of São Paulo to develop the Brazil­ there are 341 fully trained and active
the general community. The Itapeninga ian Tunable Filter Imager (BTFI), which astronomers (i.e. with a PhD) in Brazil (up
Radio Observatory (ROI; Figure 5), lo- is an innovative camera and integral field from no more than a handful some
cated in Atibaia, some 50 kilometres from spectrograph for the SOAR telescope. 40 years ago). This workforce is comple­
São Paulo, and operated by the Nacional Other long­term collaborations between mented by 313 postgraduate (Master’s
Space Research Institute (Instituto INPE and LNA on instrumentation for the and PhD) students. Thus, more than
Nacional de Pesquisas Espaciais [INPE]) OPD are also ongoing. 650 scientists are active in astronomical
is the only instrument available to all research. While there is a concentration
astronomers. This 18–90 GHz, 14-metre In the past, instrumentation develop- of astronomers in a few universities
antenna has not had a major upgrade ment at LNA was rather modest and and federal research institutes, the num­
since it was built in 1974. Access to more restricted to immediate OPD needs. But ber of groups in other places is rapidly
modern equipment would be very during the past decade much effort increasing as a result of the policy of the
much welcomed by the radio­astronomy has been invested in turning such activi­ federal government to strengthen sci-
community. ties into one of the fundamental pillars ence and higher education in less well­
of the institute. The LNA has built labora­ developed parts of the country. In conse­
tories and workshops, and provided quence, astronomy is being pursued
Instrumentation them with state-of-the-art equipment, today in 46 institutions (… and counting),
with a special emphasis on optical metro- which are widely spread across Brazil.
The desire to participate in both scientific logy and the handling of optical fibres While many of the smaller groups are part
research in astronomy and in techno­ for astronomy (see Figure 6 as an exam­ of physics or other related university
logical development has led to the imple­ ple). In collaboration with the IAG and departments, postgraduate education in
mentation of the necessary infrastruc- other university institutes, the LNA has astronomy is offered in 19 institutes.
ture to build astronomical instruments for built SIFS, a 1300-channel integral field
use at international observatories, such spectrograph (currently being commis­ There is not enough room here to char­
as SOAR. These efforts are concentrated sioned at the SOAR Telescope). It is also acterise all these institutes in detail.
at the LNA and INPE, in collaboration constructing the SOAR Telescope Echelle However, it may be worthwhile to briefly
with the universities and other scientific Spectrograph (STELES) and is planning enumerate the most important. With
institutions. a similar instrument for the OPD. The the IAG (see Figure 7), the University of
LNA was a member of the winning team São Paulo hosts the dominant research
While most of the activities in instrument in an international competition for the institute in astronomy in the country. It
development at INPE are related to fields detailed design study of the Wide Field is home to about 20% of the total work­
other than astronomy (e.g., observation of Multiple Object Spectrograph (WFMOS) force mentioned above. This is twice
The Messenger 144 – June 2011 5

6. The Organisation Bruch A. et al., Brazil’s Route to ESO Membership
Figure 7. Urania, the Muse of Astron­ (instrument development among them),
omy, from a picture window in the
come from the same sources, including
library on the former campus of the
Institute of Astronomy and Geophysics the government funding agency FINEP
of São Paulo University. (Financiadora de Estudos e Projetos),
as well as from Brazilian state funding
agencies, which normally do not fund
the operation of astronomical infrastruc­
ture. While other states also contribute,
FAPESP, the funding agency of São Paulo
state, plays a dominant role.
CNPq (Conselho Nacional de Desenvolvi­
mento Científico e Tecnológico), a branch
of the MCT, is extremely important as a
provider of stipends for students and
grants for established scientists. A similar
role is played by CAPES (Coordenação
de Aperfeiçoamento de Pessoal de Nível
Superior), a branch of the Ministry of
Education. Apart from stipends and
grants, CNPq also finances smaller scale
projects for individual scientists, scientific
meetings, etc. (as do the state agencies).
as many as the second most important, the end of the 1960s, but as the number Specific funding by the federal and state
the venerable Observatório Nacional in of active astronomers has increased, a governments, such as PRONEX (Pro­
Rio de Janeiro. Strong astronomy groups steep and continuing rise in the number grama a Núcleos de Excelência) and the
can also be found at INPE, located in of published papers has been observed Millennium Institutions (Institutos do
São José dos Campos, the Federal Uni­ (Figure 8). The role of Brazil as a signifi­ Milênio) in the past, and the current (vir­
versity of Rio de Janeiro (distributed cant producer of scientific papers was tual) National Institutes of Science and
between the Observatório do Valongo and recognised when it became a member Technology (INCT) has also greatly bene­
the Department of Physics), the Federal of Astronomy and Astrophysics, the lead­ fited Brazilian science. Two astronomy-
University of Rio Grande do Sul in Porto ing astronomical journal in Europe. related National Institutes have been cre­
Alegre and the Federal University of Rio ated: INCT-A (A for astrophysics), which
Grande do Norte in Natal. While all these Although optical and infrared observa­ focuses on preparing the astronomical
astronomy centres carry out research tional astronomy is predominant, Brazil­ community for the challenges and oppor­
in many fields, the Brazilian Centre of ian astronomy embraces a wide range tunities of the future, and INCT-E (E for
Physical Research (Centro Brasileiro de of special fields. There are at least 16 espaço [space]) which focuses on space
Pesquisas Físicas [CBPF]), Rio de Janeiro, major areas of astronomy that are being technology and astronomy from space.
which hosts the Brazilian branch of ICRA- actively pursued by astronomers in Brazil
net, focuses mainly on cosmology. and that have recently been identified Direct personnel costs are, of course, car­
in the context of a National Plan for ried by the employers, who are, in most
Administratively, the numerous astron­ Astronomy1. The relative importance of cases, the federal or the state govern­
omy groups are distributed between the various disciplines can be gauged ments. However, the private sector is also
government institutions, which are from the number of publications that they involved through private (in general, non-
directly subordinated to the federal Minis­ have generated. Table 1 gives the per­ profit) universities with research and higher
try of Science and Technology (CBPF, centages of papers by Brazilian authors education interests in astronomy.
INPE, LNA, ON), entities belonging to in refereed journals by area in 2008.
federal or state universities, and (increas­
ingly) private universities. Long­term strategic outlook
Funding
The community founded the Brazilian Brazil’s young and vigorous community
Astronomical Society (Sociedade Brazilian astronomy is largely publicly feels that it has gained an international
Brasileira de Astronomia [SAB]) in 1974. financed. Operating costs for facilities reputation as a respected player in global
The Society currently has 678 members. open to the entire community are borne astronomy. It is not seen as an accident
exclusively by the Federal Government, that Rio de Janeiro was chosen to host
As measured by the number of publi- normally through MCT research insti­ the IAU General Assembly in 2009, but
cations in refereed journals, scientific tutes. Funds for the development of rather as recognition of the achievements
productivity was all but non-existent until new projects and capital investments of Brazilian science. The community is
6 The Messenger 144 – June 2011

7. Publications in refereed journals by guaranteeing access to the future
generation of giant telescopes, i.e. the
300
European Extremely Large Telescope,
and opening up opportunities for
250
Brazilian industry to take part in its
development and construction;
200 – it provides access to ALMA, satisfying
and fostering the development of a
150 community of radio astronomers who
have not benefited from significant
100 investments similar to those made in
optical astronomy during the past three
decades;
50
– it opens up a wide range of opportuni­
ties for the participation in technological
0 development as part of the instrumen­
1970 1975 1980 1985 1990 1995 2000 2005 2009
Year tation programme for ESO telescopes.
Figure 8. Evolution of the number of publications by Among many other issues, this docu­ It is felt that the development model for
Brazilian astronomers in refereed journals over the
ment emphasises the need to maintain optical astronomy which Brazil has
past decades.
access to a competitive observational followed in the past, i.e., offering its sci­
infrastructure, on penalty of losing the entists a suite of instruments with diverse
aware that worldwide astronomy respected position gained by Brazilian characteristics on small and medium­
is characterised more than ever by inter­ astronomers. Different ways of achieving sized telescopes up to to the 8-metre-
national collaborations. Consequently, this purpose have been studied by the class Gemini giants, although with limited
success for a national community de- INCT-A and a special commission cre­ access in the case of the larger instru­
pends decisively on its participation in ated by the MCT. Based on these results ments, has lifted the astronomical com­
the international community. the broad majority of the astronomical munity to a level of maturity. This pro­
community came to the conclusion that gress now permits the next step — or
Moreover, it is understood that the grow­ the association of Brazil with ESO would rather leap — in its evolution: the ascent
ing necessity for international collabo- be the most effective of all the available to a new and higher level in scientific,
rations, the numerous scientific opportu­ options. More than any other alternative, technological and instrumental terms,
nities that present themselves in the the association with ESO benefits the which is expected to be the natural con­
worldwide scenario, combined with the country in many ways, the most impor­ sequence of Brazil’s association with
elevated costs for large-scale scientific tant advantages being that: the strongest organisation in ground­
projects, call for a medium- and long- – it gives Brazil immediate access based astronomy in the world. We are
term strategic plan for astronomy to to ESO’s existing telescopes, fostering confident that not just optical astronomy
direct and coordinate the further develop­ scientific collaboration (and competi­ will be strengthened, but that the fertile
ment of the field in Brazil. Therefore, tion!) with scientists of other member environment of partnership with ESO
with the active support of the Ministry of states, and enlarging the scope of will benefit Brazilian astronomy as a
Science and Technology, in 2010 the instruments already at the disposal of whole, as well as related technological
community elaborated a National Plan for Brazilian astronomers significantly, fields.
Astronomy1 as a guideline for the future thus eliminating some limitations felt by
of astronomy in the country, aligned to parts of the community;
the general policy for science and tech­ – it meets one of the main recommenda­ References
nology of the federal government. tions of the National Plan for Astronomy de Zeeuw, P. T. 2011, The Messenger, 143, 5
Prazeres, A. 2004, Georg Marcgrave, e o
desenvolvimento da astronomia moderna na
Optical and infrared stellar astronomy 28.8 % Table 1. Percentage of papers pub­ América Latina, na cosmopolita Recife de Nas-
Theoretical cosmology 17.4 % lished in refereed journals by area in sau, http://www.liada.net/NASSAU%20&%20
2008. GEORG%20MARCGRAVE.pdf
Optical and infrared extragalactic astronomy 11.9 %
Videira, A. A. P. 2007. História do Observatório
Physics of asteroids 5.8 % Nacional: a persistente construção de uma identi-
Theoretical stellar astrophysics 4.3 % dade científica. Río de Janeiro: Observatorio
Chemical evolution of stellar systems 4.3 % Nacional
Dynamical astronomy 4.3 %
Solar radio astronomy 3.2 % Links
Instrumentation 3.2 %
1
Exoplanets 2.7 % National Plan for Astronomy: http://www.lna.br/
PNA-FINAL.pdf
Other 13.2 %
The Messenger 144 – June 2011 7

8. Telescopes and Instrumentation
The first European ALMA antenna from
the AEM Consortium (Thales Alenia
Space, European Industrial Engineering
and MT-Mechatronics) being carried
on an ALMA transporter during the
handover to the ALMA Observatory at
the Operations Support Facility (OSF).
After testing at the OSF, it will be moved
to the ALMA Operations Site on the
Chajnantor plateau. See Announcement
ann11022 for more details.

9. Telescopes and Instrumentation
The Science Impact of HAWK-I
Ralf Siebenmorgen1 in one-hour on-source integration are: summer of 2008. The instrument also
Giovanni Carraro1 23.9 in J, 22.5 in H and 22.3 in Ks. suffered from radioactive events which
Elena Valenti1 contaminated two of the four chips of
Monika Petr-Gotzens1 The efficiency, defined as the proportion the detector mosaic (Finger, 2008). The
Gabriel Brammer1 of photons converted into electrons contamination can be seen in the dark
Enrique Garcia1 passing the telescope, instrument optics exposures. One of the four detectors
Mark Casali1 and detector, is computed for various shows on average a well­localised decay
near-infrared (NIR) instruments and is every 75 s. The event affects an area
shown in Figure 1 for the NIR cameras of 7 × 7 pixels and is eliminated by a
1
ESO SOFI, VISTA, ISAAC, CONICA and cleaning algorithm in the pipeline. Another
HAWK-I. The efficiency of the HAWK-I detector is similarly affected, and, al-
instrument is 70–80 % and so it is the though the events are much less frequent,
HAWK-I is ESO’s most efficient near- most efficient NIR camera in ESO’s instru­ they generate charge which is not local­
infrared camera, and after two and mentation suite. The stability of the zero ised to within a few pixels, but spreads in
a half years of operations we review its point is important for absolute photom- a diffuse charge cloud with an unpre-
science return and give some future etry. For HAWK-I, there is a small periodic dictable location, resulting in glitches that
directions in the context of the Adaptive scatter in the zero point of Δ J ~ 0.1 mag cannot be cleaned during data analysis.
Optics Facility. The instrument under- over a period of a year, significantly lower However, the sensitivity limit of the indi­
went major technical challenges in the than that of either CONICA or ISAAC. vidual detectors shows that there is no
early phase of its operations: there major degradation of the detection limit
was a problem with the entrance win- Along with the distortion caused by the caused by these radioactive events.
dow, which was replaced, and radio- instrument optics, atmospheric refrac-
active events occur in the material of tion produces a geometrical shrinkage of The HAWK-I instrument team has recently
two of the four detectors. A number of the field of view with increasing zenith undertaken observations to assess the
high quality science papers based on distance. The differential achromatic re- relative sensitivities of the four HAWK-I
HAWK-I data have been published, indi- fraction is ~ 0.6 arcseconds, as measured detector chips, using observations of the
cating a good performance and scien- over the full 7.5 by 7.5 arcminute field size high Galactic latitude field around the
tific return. HAWK-I is well-suited for a of HAWK-I and for a zenith distance z = 2.7 quasar B0002-422 (α 00h 04m 45s,
variety of attractive science cases and between 0° and 60°. δ -41° 56; 41?) taken during technical
a project is in development to provide time. The observations consisted of four
a faster readout, which would improve During science operations three techni- sets of 11 × 300 s sequences in the
the capabilities for Galactic observa- cal challenges were identified: the en- NB1060 filter; details of such an obser-
tions. When combined with the laser- trance window, radioactive events in the vational set-up are discussed in the
assisted ground layer adaptive optics detector material and the instrumental HAWK-I User Manual. The four sequences
system, HAWK-I will become an excel- distortion correction. The instrument was are rotated by 90° in order that a given
lent facility for challenging follow-up first installed in July 2007. At the begin­ position on the sky is observed by each
observations of exoplanetary transits. ning of the observing period P81 in 2008 of the four chips of the HAWK-I detec-
the instrument suffered from a damaged tor. The jitter sequences are reduced fol­
coating of the entrance window. This lowing the standard two-pass back­
Instrument overview and performance defect was fixed by a replacement win­ ground subtraction work flow described
dow installed during an intervention in the in the HAWK-I pipeline manual. Objects
HAWK-I is a cryogenic wide-field camera
installed at the Nasmyth A focus of the
VLT Unit Telescope 4 (UT4). The field of | |
Figure 1. Comparison
courtesy P. Hammersley
of the efficiency of the
view is 7.5 by 7.5 arcminutes, with a cross- Y J H Ks
NIR instruments SOFI,
shaped gap of 15 arcseconds between 0.8 VISTA, ISAAC, CONICA
the four 2RG 2048 × 2048 detectors. and HAWK-I is shown.
The pixel scale is 0.106 arcseconds. The
0.6
Efficiency
instrument is offered with ten filters in
two filter wheels: four broadband filters
(Y, J, H and Ks), which are identical to 0.4
the filters used in VIRCAM/VISTA, and HAWK-I
six narrowband filters (Brγ, CH4, H2, CONICA
0.2 ISAAC
1.061 μm, 1.187 μm, and 2.090 μm). The
VISTA
image quality is seeing-limited down to SOFI
at least 0.4 arcseconds. Typical limiting
magnitudes (Vega) to reach a signal-to- 1.0 1.5 2.0
noise ratio (S/N) of five on a point source Wavelength (µm)
The Messenger 144 – June 2011 9

10. Telescopes and Instrumentation Siebenmorgen R. et al., The Science Impact of HAWK-I
16 AOF and GRAAL 1. Galaxy evolution from deep multi-
CHIP 1 colour surveys;
14
CHIP 2 The Adaptive Optics Facility (AOF; see 2. Multi-wavelength observations of
12 Lelourn et al., 2010; Paufique et al., 2010 normal and active galaxies;
CHIP 3
and Arsenault et al., 2010) will provide a 3. Structure and evolution of nearby
N mag –1 arcmin – 2
10 CHIP 4 correction of the ground layer turbulence, galaxies;
Coadded stack improving the image quality of HAWK-I. 4. Galactic star and planetary formation;
8 The resulting point spread function (PSF) 5. Outer Solar System bodies.
6
diameter that collects 50 % encircled
energy is reduced by 21% in the Ks­band, HAWK-I started to operate regularly in
4 and by 11% in the Y­band, under median April 2008. A significant number of
seeing conditions at Paranal of 0.0.87 arc- observations executed during P81 were
2 seconds at 500 nm. Hence, the AOF will affected by the damaged entrance
0
provide better seeing statistics. When window coating, and were re­executed
13 14 15 16 17 18 19 installing the AOF on UT4, the secondary by ESO. In the period from mid-2008
MAG_APER (D = 1.8 , ZP = 25) mirror of the telescope will be replaced until end of 2010, 26 refereed papers
by a deformable secondary mirror (DSM) were published containing HAWK-I
Figure 2. Number counts as a function of aperture with more than 1000 actuators. In addi­ results. They have 350 citations to date
magnitude of the four HAWK-I detectors: chip1 (red
tion, four laser guide stars will be installed and an h-index of 10. Of these 26
line), chip2 (orange), chip3 (green), chip4 (blue line)
and the co-addition of all four chips (black line). on the telescope structure, and a wave­ papers, two were published in Nature,
Dashed lines give the number of spurious detections. front sensor system, GRAAL (ground seven in ApJ, four in ApJ Letters, one in
Radioactive events are most common for chip 2, layer adaptive optics assisted by lasers), AJ, four in MNRAS, and 12 in A&A. The
which nevertheless has a similar detection probabil­
will be used to measure the turbulence two Nature papers, Tanvir et al. (2009)
ity as the other chips, but an enhanced number of
spurious detections at faint flux (> 17 mag) levels from artificial guide stars. GRAAL will be and Hayes et al. (2010), resulted in ESO
(shown as dashed orange). installed between HAWK-I and the press releases. To evaluate the science
Nasmyth flange. HAWK-I’s field of view is impact of HAWK-I, we have compared
not affected by GRAAL. It is planned to the number of papers based on data
begin installing the AOF in 2013, with a obtained at the other NIR VLT instruments
total telescope downtime of a few months during their first 2.5 years of science
(subject to the exact distribution of tech­ operations. The rate of publication turns
are detected using the SExtractor soft­ nical time) due to the installation of the out to be fairly similar among all the
ware. The resulting number counts as new secondary mirror, the lasers and VLT instruments considered (NACO,
a function of aperture magnitude ob- GRAAL. The schedule anticipates that ISAAC, SINFONI and CRIRES). The sci­
served by each detector are shown in the AOF will be operational from 2015. ence output of HAWK-I up to the end
Figure 2. The limiting magnitudes, here of 2010 can be summarised as follows:
taken to be the magnitude where the Normal adaptive optics systems aim 1) In most cases, publications which are
number counts in Figure 2 decrease at correcting atmospheric turbulence based on HAWK-I present results on
sharply, provide a proxy for the individual down to the diffraction limit of the tele­ extragalactic, high redshift astrophys­
detector sensitivities. The sensitivities scope. The price to be paid is a limit in ics. The most relevant papers being
agree to within 10 % between the individ­ corrected field of view (less than 1 arc­ the characterisation of the galaxy pop­
ual chips. We also show in Figure 2 minute) and a limit in sky coverage (less ulations around z ~ 2 (Galametz et al.,
the number counts for a deep co-added than 50 %) since a bright guide star is 2010; Hayes et al., 2010; and Lidman
stacked image of the four rotated and required even when using laser guide et al., 2008) and beyond redshift z ~ 6
aligned jitter sequences which are a fac­ stars. The AOF ground layer adaptive (Vanzella et al., 2010; Fontana et al.,
tor of two deeper than the individual optics mode (GLAO) does not provide 2010; Castellano et al., 2010a,b; and
sequences. We used the co-added stack diffraction-limited image quality, but it Bouwens et al., 2010). Such a burst
to assess the number of spurious does correct the full 7.5 by 7.5 arcminute of results for extremely high redshift
sources detected on each individual de­ field of view and the sky coverage is targets was not expected at the time
tector: objects matched from the single practically 100 %. when defining the HAWK-I science
chip image to the deeper image are cases, while the results at intermediate
considered to be real, while objects that redshifts were expected from science
only appear on the single chip images HAWK-I science return case #1.
are considered spurious. The image arte­ 2) The other fields explored so far are
facts on detector 2, which are caused When HAWK-I was conceived, the Milky Way stellar populations (Brasseur
by radioactive events, do result in an ele­ selected science cases, according to et al., 2010), trans-Neptunian objects
vated number of spurious detections the document, Science Case for (Snodgrass et al., 2010), gamma-ray
at faint magnitudes, reaching 20 % at the 0.9–2.5 μm infrared imaging with the bursts (D’Avanzo et al., 2010) and qua­
limiting magnitude. VLT (ESO/STC-323), were: sars (Letawe & Magain, 2010). Stellar
10 The Messenger 144 – June 2011

11. population studies have been ham­ Figure 3. Three colour
ESO/M. Gieles, Acknowledgement: Mischa Schirmer
(J [1.25 µm], H [1.65 µm]
pered by HAWK-I’s large minimum
and Ks [2.15 µm]) com­
detector integration time (DIT), which posite maps obtained
causes saturation on bright sources with HAWK-I. The upper
and almost completely prevents ob- image shows the nearby
galaxy Messier 83, total
servations in the Galactic disc.
exposure time was 8.5
3) No papers were published in the field hours and field of view
of star formation and structure of 13 arcminutes squared.
nearby galaxies (science cases #3 and On the bottom, an
image of 6 by 5.2 arc­
#4) in the period up to and including
minutes of two stellar
2010, in spite of the fact that several clusters in the Carina
programmes have been queued and Nebula is shown,
successfully executed. obtained during HAWK-I
science verification.
4) Contrary to expectations, HAWK-I was
intensively used to study exoplanets,
via transit or occultation techniques
(Gibson et al., 2010; Anderson et al.,
2010; and Gillon et al., 2009), and to
conduct supernova search campaigns
(Goobar et al., 2009) for spectroscopic
follow-up. Transit observations, in
particular, are expected to be increas­
ingly important in the nearby future
as a windowed readout of the detec­
tors has been implemented.
5) The majority of the observations pub­
lished require or benefit from the large
field of the instrument.
In Figure 3 we give two examples of JHKs
colour-composite maps highlighting
the superb image quality of the HAWK-I
camera.
Future directions: HAWK-I + GRAAL
It is anticipated that HAWK-I will be
equipped with GRAAL and routinely op-
erate in GLAO mode from 2015 onwards,
which will open up new paths for
competitive science cases in the coming
years. The image quality delivered by
HAWK-I + GRAAL is expected to be 20 %
better in comparison with today. For
seeing in the Ks-band of 0.6 arcseconds,
the GRAAL-supported instrument is
expected to deliver a resolution of mum, FWHM) on point sources larger good seeing conditions for NACO and
0.5 arcseconds on a regular basis. Given than 0.6 arcseconds. This arises from the SINFONI. Observing with HAWK-I
HAWK-I’s pixel scale of 0.106 arcseconds, fact that 70 % of the HAWK-I observa­ together GRAAL will result in a much
the PSF delivered by HAWK-I + GRAAL tions were executed during DIMM (differ- better image quality performance. The
will still be Nyquist-sampled, which ential image motion monitor) seeing question arises: what kind of scientific
is particularly important for precise PSF- worse than 0.83 arcseconds. Half of the projects will be feasible with HAWK-I +
photometry, astrometry and the analysis HAWK-I observations were performed at GRAAL that are currently not feasible
of morphological structures on sub- a median DIMM seeing of almost 1 arc­ with HAWK-I, or only under very rare con­
arcsecond spatial scales. Currently, half second. The poorer than average seeing ditions, when the seeing is exceptionally
of the HAWK-I Ks­band images show conditions prevalent during most HAWK-I good. We outline three selected science
an image quality (full width at half maxi­ observations is a result of the demand for cases of HAWK-I + GRAAL.
The Messenger 144 – June 2011 11

12. Telescopes and Instrumentation Siebenmorgen R. et al., The Science Impact of HAWK-I
1. Cosmological surveys HAWK-I instrument operation team is at to be followed up around stars signifi­
A deep, wide-field NIR imaging survey present testing a new windowed detec- cantly fainter than those observed at the
complementing the HST/CANDELS cos­ tor readout scheme that allows very short moment (Ks of 8–11 mag). Therefore a
mological survey is required. CANDELS1 exposure times on the brightest pixels larger volume of planet–host star systems
is the largest single project in the history and, in parallel, long exposures for the re- can be probed, so that potential exoplan­
of the Hubble Space Telescope, with maining field. Such a new detector read­ ets detected by CoRoT come within
902 assigned orbits of observing time out mode in combination with HAWK-I + reach of the VLT and hence provide
and obtains images at J­ and H­band GRAAL’s improved seeing capabilities important NIR constraints on the physical
over a total field of view of 30 × 30 arc­ should lead to an increase of HAWK-I ob- nature of the planets. Observations with
minutes. The survey will be completed in servations in this research field. VISTA will not have the required sensi-
2014. As the scientific exploitation also tivity to perform such investigations.
relies on multi-colour imaging, HAWK-I + Since the large field of view is important
GRAAL is an ideal instrument to com- 3. Exoplanets and transits for precision photometry, there is no
plement the survey with very deep Ks­ HAWK-I has recently proved to be an strong advantage in using JWST/NIRCam
and Y­band imaging, as well as with nar­ excellent instrument with which to instead.
rowband imaging aimed at searching perform challenging observations of exo­
for very high redshift galaxies. Morpho­ planetary transits. In order to obtain an
logical studies of galaxies at intermediate overall picture of an exoplanet’s atmos­ References
and high z are a particular goal of the pheric properties, occultation data in Anderson, D. R. et al. 2010, A&A, 513, 3
project that can be pursued only with a many photometric bands are required. Arsenault, R. et al. 2010, The Messenger, 142, 12
spatial resolution of < 0.5 arcseconds With a continuously growing number of Bakos, G. A. et al. 2011, AAS Meeting 217, 253.02
over a wide area. A wide field of view is newly discovered planets and planetary Bouwens, R. J. et al. 2010, ApJ 725, 1587
Brasseur, C. A. et al. 2010, AJ, 140, 1672
essential in such a study, since structural candidates, there is a high demand for Cameron, A. C. et al. 2009, IAU Symposium,
properties are analysed on sufficiently comprehensive follow-up observations by Volume 253, 29
large statistical samples. HAWK-I obser­ NIR imaging. Crucial requirements for Castellano, M. et al. 2010a, A&A, 511, 20
vations in the Y-band, complementing such observations are a wide field of Castellano, M. et al. 2010b, A&A, 524, 28
Coppin, K. E. K. et al. 2010, MNRAS, 407, L103
the first two CANDELS fields, have already view, allowing for a large number of refer­ D’Avanzo, P. et al. 2010, A&A, 422, 20
been scheduled. ence sources for precise relative photom­ Decarli, R. et al. 2009, ApJ, 703, L76
etry, and an instrument sensitive enough Finger, G. Reports on HAWK-I detectors available at:
VISTA does offer the requested wide-field to collect a sufficient number of photons, http://www.eso.org/~gfinger/marseille_08/AS08-
AS12-9_H2RG_mosaic_gfi_final.pdf
capability, but delivers neither the spa- typically for a S/N > 1000, in a short time. http://www.eso.org/~gfinger/hawaii_1Kx1K/cross­
tial resolution nor the required sensitivity. From space the CoRoT satellite (Moutou talk_rock/crosstalk.pdf
In order to reach the same limiting mag­ et al., 2008) is a mission particularly Fontana, A. et al. 2010, ApJL, i725, 205
nitude, VISTA requires an integration time designed to discover transiting Galametz, A. et al. 2010, A&A, 522, 58
Gibson, N. P. et al. 2010, MNRAS, 404, L104
16 times longer than HAWK-I + GRAAL. exoplanets. CoRoT has already found Gillon, M. et al. 2009, A&A, 506, 359
However, the NIRCAM2 instrument several hundred systems with candidate Gogus, E. et al. 2010, ApJ, 718, 331
onboard JWST will have a field of view transiting planets. The mission will Goobar, A. et al. 2009, A&A, 507, 71
almost six times smaller than HAWK-I, but continue beyond 2015 and will possibly Greiner, J. et al. 2009, ApJ, 693, 1610
Hayes, M. et al. 2010, Nature, 464, 562
will offer at least a factor 15 in improved be followed up by PLATO (Roxburgh Hayes, M. et al. 2010, A&A, 509, L5
sensitivity. JWST is expected to become & Catala, 2006), an ESA project study Hickey, S. et al. 2010, MNRAS, 404, 212
operational in ~ 2016. due to be launched in 2018. Similarly, Le Louarn, M. et al. 2010, SPIE, 7736, 111
from the ground, there are robotic search Letawe, G. & Magain, P. 2010, A&A, 515, 84
Lidman, C. et al. 2008, A&A, 489, 981
projects ongoing on small telescopes. Mattila, S. et al. 2008, ApJ, 688, L91
2. Nearby wide-field imaging Instrumentation includes wide-field imag­ McLure, R. J. et al. 2010, MNRAS, 403, 960
Stellar population studies, both in nearby ing capabilities covering several degrees Moutou, C. et al. 2008, A&A, 488, L47
galaxies and in Galactic fields, currently in optical bands. The goal is to discover Paufique, J. et al. 2010, SPIE, 7736, 57
Roxburgh, I. W. & Catala C. 2006, IAUJD, 17, 32
suffer most from crowding and will bene­ a large sample of candidate planetary Snodgrass, C. et al. 2010, A&A, 511, 72
fit from an improved Ks image quality transits which will be followed up on Stanishev, V. et al. 2009, A&A, 507, 61
provided by HAWK-I + GRAAL. High spa­ larger telescopes by radial velocity stud­ Tanvir, N. R. et al. 2009, Nature, 461, 1254
tial resolution coupled with a wide field ies or NIR imaging. Examples are: WASP Vanzella, L. et al. 2010, ApJL, 730, 35
of view is an important requirement for (Cameron et al., 2009) which has already
stellar population studies. A problem of detected 16 systems and will continue Links
current HAWK-I observations, when tar­ for several years; or HAT-South, which is
1
geting crowded stellar populations, is that the first global network dedicated to CANDELS: www.candels.ucolick.org
2
JWST NIRCam:
the relatively large minimum DIT of 1.7 s search for transiting planets. www.ircamera.as.arizona.edu/nircam
causes saturation on the brightest
sources, which are numerous when ob­ The increase in sensitivity of HAWK-I +
serving towards the Galactic disc. The GRAAL will allow exoplanetary transits
12 The Messenger 144 – June 2011

13. Telescopes and Instrumentation
p3d — A Data Reduction Tool for the Integral-field
Modes of VIMOS and FLAMES
Christer Sandin1 Feature ESO pipelines p3d Table 1. Comparison
between features of p3d
Peter Weilbacher1 Logging, at different levels of verbosity x x
and the IFU modes of
Ole Streicher1 Configuration by a plain text file x x
the ESO VIMOS (version
Carl Jakob Walcher1 Combination of raw-data images partly all recipes 6.2) and FLAMES (ver­
Martin Matthias Roth1 Dark current subtraction x – sion 2.8.7) pipelines.
Spectrum extraction:
regular/deconvolution methods x/– x/2
1 Spectrum extraction:
Leibniz-Institut für Astrophysik Potsdam
subtraction of a scattered-light component – x
(AIP), Germany
Fully automatic spectrum tracing x x
Creation of a dispersion mask automatic interactive
Flat-field normalisation partly x
The second release of the data reduc-
Flux calibration x –
tion tool p3d now also supports the
Full error propagation partly x
integral-field modes of the ESO VLT
Interactive inspection of intermediate and
instruments VIMOS and FLAMES. final products – x
This article describes the general capa- Reduction using a GUI/scripts x/x x/x
bilities of p3d and how its different
tools can be invoked, with particular
reference to its use with data from using the DCR program (Pych, 2004) first, solution to the problem. p3d comes with
VIMOS and FLAMES. and thereafter, if required, combining an integrated spectrum viewer that
the resulting images in p3d using an works with any IFU (row-stacked) spec­
average. All extracted images of p3d are trum image, together with a fibre position
p3d is a general and highly automated accompanied by an error image. table.
data reduction tool for fibre-fed integral
field unit (IFU) spectrographs. Based By default p3d shows graphical results The algorithms used in p3d are described
on an early proprietary version, p3d was of the spectrum tracing, the cross- in Sandin et al. (2010). With this new
rewritten from scratch to be more ver- dispersion profile fits (used later when release all parts of p3d are now thoroughly
satile, user-friendly, extendable and deconvolving overlapping spectra), the documented. The installation procedure
informative (Sandin et al., 2010). The first quality of the dispersion solution, and is described in the distribution README
release supported four IFUs: the lens the optimally ex tracted spectra. Figure 1 file, and the various recipes are, together
array and PPAK of the PMAS spectro­ shows an example. This makes it easy with all the options, described in detail
graph at the Calar Alto Observatory; to check that the outcome is correct and in the headers of the respective files. A
SPIRAL at the AAOmega spectrograph satisfactory; and if it is not these plots more appealing version of the same doc­
at the Australian Astronomical Observa­ will quickly provide important clues for a umentation is available at the project
tory; and VIRUS-P at the McDonald
Observatory. The second release of p3d
supports most of the remaining instru­
ments, including the four higher resolu­
tion IFU modes of VIMOS (HR-Blue,
HR-Orange, HR-Red, and MR), as well as
all the setups for the three IFU modes
of FLAMES (ARGUS, and the two sets of
mini IFUs).
Data reduction features
All the reduction capabilities of p3d, with
supporting test studies, are described
in detail in Sandin et al. (2010). p3d itself
is available at the project website1. In
Table 1 we outline the available features
of p3d and the two ESO pipelines for Figure 1. The fitted cross-dispersion line profiles for
a set of the spectra in the VIMOS fourth quadrant
VIMOS (version 6.2) and FLAMES (i.e.
(with grism HR-orange). The different lines are: inten­
GIRAFFE; version 2.8.7). Cosmic-ray hits sity (in raw counts) at the middle column of the bias-
in single images, or in images that cannot subtracted continuum image (black line); the fitted
be combined, are not removed by p3d. Gaussian profiles (blue lines); the initial position of
each spectrum (vertical red lines); and the vignetted
Instead, for ESO data, we recommend
spectra, which were not fitted (vertical blue lines).
The Messenger 144 – June 2011 13

14. Telescopes and Instrumentation Sandin C. et al., p3d — A Data Reduction Tool
<ob900000.sh> <ob900000.pro>
#!/bin/bash
cpath=`pwd` cd,cur=cpath
path=”/data/user/VLT-P87/C/2011-04-27” path=’/data/user/VLT-P87/C/2011-04-27’
cd $path cd,cpath
name=”ngc1-hr-blue-T1-1a” name=’ngc1-hr-blue-T1-1a’
parfile=”${p3d_path}/data/instruments/vimos/nvimos_hr.prm” parfile=!p3d_path+’/data/instruments/vimos/nvimos_hr.prm’
userparfile=”../p3dred/user_p3d.prm” userparfile=’../p3dred/user_p3d.prm’
opath=”../p3dred/odata/$name” opath=’../p3dred/odata/’+name
mkdir -p $opath file_mkdir,opath
df1=” df1=[, $
VIMOS_IFU_OBS117_0001_B.1.fits.gz, ‘VIMOS_IFU_OBS117_0001_B.1.fits.gz’, $
VIMOS_IFU_OBS117_0002_B.1.fits.gz, ‘VIMOS_IFU_OBS117_0002_B.1.fits.gz’, $
VIMOS_IFU_OBS117_0003_B.1.fits.gz, ‘VIMOS_IFU_OBS117_0003_B.1.fits.gz’, $
VIMOS_IFU_OBS117_0004_B.1.fits.gz” ‘VIMOS_IFU_OBS117_0004_B.1.fits.gz’]
group=1,1,1,2 # Files 1-3 are combined, file 4 is used single group=[1,1,1,2] ; Files 1-3 are combined, file 4 is used single
# Extracting the object spectra for quadrant 1: ; Extracting the object spectra for quadrant 1:
logfile=”../p3dred/logs/dred_${name}_objx_q1.log” logfile=’../p3dred/logs/dred_’+name+’_objx_q1.log’
masterbias=”../p3dred/odata/VIMOS_SPEC_BIAS118_0001_B_mbias1.fits.gz” masterbias=’../p3dred/odata/VIMOS_SPEC_BIAS118_0001_B_mbias1.fits.gz’
tracemask=”${opath}/VIMOS_IFU_LAMP118_0001_B_imcmb1_trace1.fits.gz” tracemask=opath+’/VIMOS_IFU_LAMP118_0001_B_imcmb1_trace1.fits.gz’
dispmask=”${opath}/VIMOS_IFU_WAVE118_0001_B.1_dmask1.fits.gz” dispmask=opath+’/VIMOS_IFU_WAVE118_0001_B.1_dmask1.fits.gz’
flatfield=”${opath}/VIMOS_IFU_LAMP118_0001_B_imcmb1_flatf1.fits.gz” flatfield=opath+’/VIMOS_IFU_LAMP118_0001_B_imcmb1_flatf1.fits.gz’
${p3d_path}/vm/p3d_cobjex_vm.sh $df1 $parfile masterbias=$masterbias p3d_cobjex,df1,parfile,masterbias=masterbias, $
tracemask=$tracemask dispmask=$dispmask flatfield=$flatfield tracemask=tracemask,dispmask=dispmask,flatfield=flatfield, $
userparfile=$userparfile opath=$opath detector=0 userparfile=userparfile,opath=opath,detector=0, $
logfile=$logfile loglevel=2 group=$group & logfile=logfile,loglevel=2,group=group
# Click away the popup window (for a 1600x1200 screen):
sleep 1 && xdotool mousemove 800 600 && xdotool click 1
Figure 2. An example of a script that can be used to
extract object spectra in VIMOS data. The script on
the left-hand (right-hand) side is used from the shell
(IDL command line).
website1; these web pages are updated after any change to the procedure or the are all traced well, without any required
with each new release. code. Figure 2 shows an example of a user interaction. The third quadrant
simple script, using both methods, which sometimes requires a manual parameter
p3d is based on the Interactive Data can be used to reduce VIMOS data. adjustment to trace all the spectra
Language (IDL)2, which must be installed properly; this is caused by the spectrum
on the system. All computing platforms pattern, which is less well defined than
supported by IDL can be used with p3d. Details regarding VIMOS and FLAMES in the other quadrants. The tracing plots
There are three ways to invoke p3d. The show that the tracing procedure some­
first is through the graphical user inter­ When p3d is used with FLAMES and times misses one spectrum in the last
face (GUI), which can be started either VIMOS some care is required in the group of spectra. With pre-refurbishment
from the IDL command line or using the configuration procedure to produce the data, a similar problem is only found in
shell script provided. This approach cor­ most accurate outcome possible. We data from the fourth quadrant. The scat­
responds to the ESO tool Gasgano. The emphasise that the required modifica­ tered­light subtraction should be used
second is to run the individual recipes tions are small when comparing data in all spectrum extraction procedures to
from the command line, and the third is that were extracted either before or set the zero background level properly;
to use the shell scripts provided; this last after the respective refurbishments (cf. we recommend a zeroth-order polyno­
approach most closely corresponds to Hammersley et al., 2010; Melo et al., mial fit.
the ESO tool Esorex. The shell scripts use 2007). Here we note the details of each
the IDL Virtual Machine together with the instrument separately, beginning with We found that the first-guess dispersion
compiled binary files that are provided, VIMOS. solution of p3d allows the emission lines
with or without an IDL license. The shell that are required to create an accurate
scripts work on all platforms with a bash VIMOS dispersion mask to be easily identified
shell. With VIMOS data the reduction is done for all grism setups and quadrants. For
for each of the four quadrants individually. our data from P86 (PI: Lundqvist), the
The GUI method is an easy entry point The data from the four quadrants are maximum residual (for HR-blue and HR-
for the new user. By comparison, the two combined in a final step — after the data orange) between the true wavelength
script methods allow the more experi­ have been flux calibrated — to produce and the fitted wavelength of any arc line
enced user to save time, since she or he a datacube image with all 1600 spectra. was 0.002–0.007 nm for a fifth-order
can simply execute the scripts anew, Data from quadrants one, two and four, polynomial. Larger residuals are found in
14 The Messenger 144 – June 2011

15. Figure 3. The p3d spectrum viewer showing an
extracted datacube, where all four quadrants of
VIMOS have been combined. The four different pan­
els show: the spectrum image (upper left); the spatial
map at a selected wavelength (upper right; north is
up and east left); ten stored spatial maps of different
wavelengths (middle panels); the selected spectrum,
in this case the average of the 33 spectra that are
marked in the spatial map (bottom panel).
low-transmission spectra. We also found The data were not flux-calibrated, but the strict the set of arc lines to the brightest
that the highest accuracy level can be data from the separate quadrants were before the reduction is begun. In our
achieved in more spectra if cosmic-ray re-normalised using the mean flat-field data from P83 (programme ID 083.B-
hits are removed in the arc image before spectrum of each quadrant. 0279, PI: Neumayer), the maximum resid­
creating the dispersion mask. ual, between the true wavelength and
FLAMES the fitted wavelength of any arc line, is
Noise reduction is a good reason to The three different IFU modes of FLAMES constant at about 0.005–0.006 nm, for a
replace an extracted flat-field image with use the same instrument configuration fourth-order polynomial using about 20
a smoothed version. Such a replace- file. Since there is only one detector, all lines and the LR02 setup. While the fring­
ment proved impossible with VIMOS, due spectra are reduced at once. We have ing effects in the red wavelength range
to the strong fringing at red wavelengths. found that the tracing works well in all are lower with the refurbished instrument
With the new data the fringing effects cases, although the last sky fibre is than previously, one should still not
are smaller, but still present. The default always outside the CCD. The calibration smooth the flat-field data to remove the
is therefore to avoid any smoothing of the fibres are reduced along with the other fringes more completely.
flat-field image. Moreover, if twilight flat- fibres, but are never used by p3d. Fur­
field images are available, it is possible to thermore, p3d provides a linear first- Our reduced data of the nuclear region
use their transmission correction and guess dispersion solution for the same of the galaxy NGC 3621 were fitted with
correct the data further. In Figure 3 we set of arc lines that is used by the stellar population models and are shown
show the spectrum viewer display for an GIRAFFE pipeline. However, in order to in Figure 4; specifically we used the pixel-
extracted and combined dataset of a enable easy identification of all the arc fitting code PARADISE (Walcher et al.,
supernova remnant (using HR-orange). lines to be used, it is advisable to re­ 2009), as well as a preliminary version of
The Messenger 144 – June 2011 15