THE ATACAMA LARGE MILLIMETER ARRAY (ALMA)

A NEXT GENERATION RADIO TELESCOPE

Jacob W.M. Baars

European Southern Observatory

Karl-Schwarzschild-Strasse 2, 85748 Garching, Gennany
[email protected]

ABSTRACT cosmos are at such temperatures. Also many

We describe in general terms the main characteristics interstellar molecules emit spectral lines in this
of a new, very large Radio Telescope for the wavelength region, enabling us to study the chemistry
(sub )millimeter wavelength region. It is an aperture and dynamics of such molecular clouds. The infrared
synthesis array telescope consisting of 64 individual radiation from galaxies at very large distances arrives
reflector antennas of 12 m diameter, which are at earth shifted to mm-wavelengths, giving us the
interconnected to form 2016 simultaneous two- possibility to study galaxy formation in its very early
element interferometers. The total collecting area is stages, when the Universe was young.
more than 7000 m2 and the angular resolution will be
variable from ~ 10 arc seconds to ~ 10 milliarcseconds. Radio astronomy has long suffered from its poor
The telescope covers the entire frequency range from angular resolution with respect to optical astronomy.
30 - 950 GHz with 10 separate receiver units. To This was partially overcome in the early days by using
minimize atmospheric effects, it will be located at interferometry. But only the development of aperture
5000 m altitude in Northern Chile. The project is a synthesis combined imaging capability with high
collaboration between North America (USA and resolution. The largest telescope of this class (the
Canada) and Europe with the likely later participation Very Large Array in the USA) achieves a resolution
of Japan. After several years of development, the of better than 1 arcsecond, comparable to that of earth-
project is now entering the construction phase. bound optical telescopes. Present day astrophysics
Completion is scheduled for 2011. requires the combination of observations in the entire
range of the available electromagnetic spectrum at
comparable resolution. New telescopes with adaptive
INTRODUCTION optics (ESO-VLT) or located in space (HST) achieve
In astronomy the wavelength region between O.l and an angular resolution well below 1 arc second. The
3 mm, the millimeter and submillimeter range, is Atacama Large Millimeter Array (ALMA), to be
increasingly being exploited, but its instrumental described in the sequel, will achieve this in the region
capabilities still are behind the state in other of the (sub )millimeter wavelengths, covering all
wavelength regimes. There are two principal reasons accessible atmospheric "windows" from 30 - 950
for this situation: firstly, technology of antennas and GHz. ALMA is the result of merging two proposals
electronics is extremely difficult in this frequency for mm-arrays, one from Europe and the other from
range and moreover lacks commercial application; the USA. The combined instrument will provide
secondly, the earth's troposphere (in particular water astronomers with an order of magnitude increase in
vapour) strongly absorbs mm-radiation. To be sensitivity, angular resolution and spectral coverage
effective, mm-telescopes must be located at high and compared to existing mm-telescopes. The technical
dry sites which are difficult and expensive to operate. specifications of ALMA are assembled in Table 1.

On the other hand, radiation from astronomical objects

in this wavelength regime are of great importance for ARRAY CONFIGURATION AND SITE
the understanding of several key astrophysical ALMA will be an aperture synthesis array of 64
problems, in particular the formation of stars and transportable reflector antennas of 12 m diameter
galaxies. (Sub )millimeter radiation originates in areas each. Connecting these antennas in pairs delivers 2016
of low temperature (10-1 OOK) and vast regions of the simultaneous baselines of varying orientation and
length. Thus we sample 2016 points in the synthesized reflectors but with an angular resolution proportional
aperture, also called the uv-plane. By tracking the to the largest baseline in the synthesized aperture.
object for a certain time the projected baseline rotates
with respect to the sky and hence samples additional ALMA is conceived as a "zoom" telescope with a
points in the uv-plane. After combining all widely varying angular resolution. The compact
measurements and applying a Fourier Transformation configuration functions as an almost filled aperture
we obtain the brightness distribution of the area of and provides a very high brightness sensitivity with
sky, bounded by the antenna beam of the individual moderate angular resolution. The resolution can be
improved in steps of roughly three by moving the
antennas away from each other (see Table 1) up to a
Table 1: ALMA technical specification summary largest baseline of about 14 km, where the highest
Array number of antennas 64 resolution will be about 5 milli-arcseconds at 0.32 mm
total absorption area 7200 m2 wavelength. To enable this flexibility, the antennas are
angular resolution (arcsec)0.2 A(mm)/B(km) transportable and will be moved by a large transporter
to any of a total of about 250 stations, equipped with
Configuration foundations and cable connections.
compact 150 m diameter
zooming capability 450-5000 m The site of ALMA is the Llano de Chajnantor in the
Atacama Desert of northern Chile at _23 latitude.
0

high resolution 14km

baseline stations about 250 This rather flat plateau at 5000m altitude provides
space for the largest baselines and shows excellent
Antennas paraboloidal Cassegrain reflector atmospheric conditions for submillimeter astronomy.
diameter 12 m It is one of driest areas in the world. Access is easy on
surface accuracy 20-25 11m rms a paved road from the harbour city of Antofagasta to
pointing (in 9 m/s wind) 0.6 " rms northern Argentina, which runs close to the site.
residual pathlength error 15 11m ALMA will be remotely operated from an Operations
primary flD-ratio 0040 Support Facility at about 2800 m altitude near the
Cassegrain f/D 8.0 village of San Pedro de Atacama, which is the oldest
"fast motion" capability 1S in l.5 sec continuously inhabited village in Chile and an
transportable between stations increasingly favourite goal of tourists. The photo
below gives an impression of the site.
Receivers: SIS mixer or HFET amplifier
all dual-polarisation systems
"baseline" project 84-119 GHz
211-275 GHz
275-370 GHz
602-720 GHz
later "extensions" 787-950 GHz
385-500 GHz
125-163 GHz
31-45 GHz
163-211 GHz
67- 90 GHz
THE REFLECTOR ANTENNAS
Correlator and IF section
The selection of the diameter of the antennas, which
IF bandwidth 8 GHz
are to serve as the elements of the array, is a process
sampling 4 Gbit/sec
of compromise between technical limitations and
transmission digital
astronomical requirements. For a given total collecting
correlated baselines 2016
area, a large number of small elements provides a
spectral channels per IF 4096
larger field of view (the element beam width) and
more baselines, hence a better coverage of the uv- Because of the high demands on the performance of
plane. This is the preferred solution for the the antennas, which have not been demonstrated
observation of large molecular clouds in our Galaxy. before, and the need to acquire a set of 64 antennas in
However the complexity of the electronics system an economical way, the North American and
increases strongly. If the main interest is in the study European partners in ALMA decided to each contract
of relatively small extragalactic sources, which will be for the fabrication of a prototype antenna. A total of 9
inherently weak, the use of a smaller number of large bids were received. In the USA the proposal by
elements is advantageous. The global aspects of VertexRSI was selected, while the proposal by the
ALMA and its users led to the compromise of 64 Consortium Alcatel/EIE/Costamasnaga was selected
element antennas of 12 m diameter. by Europe. These antennas are now under fabrication.
They will be erected next to each other at the site of
The requirements on the antennas are extremely high. the VLA in New Mexico, where they will be subjected
For an acceptable performance the reflector surface to an extensive series of evaluation tests by a joint
must be accurate to ~ 10116 at the shortest wavelength. team of scientists and engineers from both partners.
Thus we impose a specification of 25 11m rms (goal 20
11m) for the surface accuracy of the antennas under It will be clear to the reader that details of the designs
operating conditions. It includes deformation caused of these antennas cannot be divulged at the present
by temperature differences in the structure, gravity time. Both designs incorporate a base and yoke of
and wind forces in a 9 m/s wind. Such accuracies can steel, while the elevation structure, with the receiver
be achieved only by the use of active control of the cabin, and the support structure for the surface panels
surface or by special materials with a small coefficient are made of CFRP. The support structure for the
of thermal expansion. An example of such a material subreflector is also of CFRP. The surface panels use
is carbon fiber reinforced plastic (CFRP). different technologies; machined aluminium plates
and an electroformed nickel replication technology,
Radio telescopes must rely on "dead pointing" (there respectively. A rendering of the European proposal is
are no "guide stars"). The beam width of a 12 m shown here to give an impression of the structure.
antenna at 0.33 mm wavelength is only 6 ". Thus we
require a pointing accuracy and stability of 0.6"
during observations in all operational conditions. To
achieve this in an economical way poses a great
challenge to the designers. In an interferometer it is
essential that the relative path lengths of the radiation
to the two elements are well known. Therefore, we
also have a specification on path length stability of 15
11m, a stringent requirement for a 12 m large structure.

The varying water vapour content of the troposphere

not only causes an absorption of the radiation, it also
creates a phase variation in the passing wave front.
These are indistinguishable from a phase effect which
is inherent to the source brightness distribution. It is
essential that these atmospheric effects be calibrated
as well as possible. In many instances this can be done
quite well by switching regularly between the source
under observation and a nearby calibration point
source with known, constant, phase behaviour. The
ALMA antennas will be capable of doing this with a
minimum loss of time: the antenna will move 1S on
the sky in a time of ~1.5 seconds. The resulting
requirements on the drive system are quite severe.
The antennas have pickup attachments for the
transporter to move them between stations according
to the selected baseline configuration. It is planned to
assemble the entire antenna, complete with all
electronic equipment, at the lower operations center
and transport the functional antenna to its position at
the 5000 m high site. For major repairs or overhaul,
the antenna will be transported down to the operations
center to avoid extensive work at the high altitude.

THE RECEIVER SYSTEM

While the antenna is the most expensive single item of
ALMA, the receiver system is by far the most
complicated. Apart from the frequency regions where
the atmosphere absorbs strongly, the entire frequency
range from 30 to 950 GHz is to be covered. Technical
limitations, in particular waveguide bandwidths, have
resulted in the adoption of ten separate receiver units, The IF bandwidth will be 8 GHz, significantly wider
which we call "frontend" (Table 1). Contrary to the than current systems. The central local oscillator
antenna, the technology for (sub)millimeter receivers might be based on a photonic system, whereby the
is not available in industry. In fact, the development of submillimeter LO-signal is obtained by mixing two
low-noise amplifier and mixer technology for this optical laser signals in a photodiode. This method has
frequency range has almost entirely been the work of been successfully applied to frequencies near 300
engineers working at radio observatories. A particular GHz and if extension to 950 GHz is achieved, the LO-
challenge is the design and construction of a huge system will be simpler and cheaper than the usual
number of receivers. Current mm-interferometers are solution with frequency multipliers.
limited to 6-10 antennas, each equipped with one or
two frontends. ALMA will require 64xlOx2 = 1280 All signals will be digitised at the antenna and sent to
units in total. The approach to this is the subdivision the central station, where a large digital correlator will
of the receiver system in a number of fully self- form the cross-correlation of the individual
contained modules. Thus the frontends for the interferometers. For the study of spectral lines, the
different bands will be developed and built at the radio correlator will provide a resolution of 4096 spectral
astronomy laboratories of the participating institutes. points over the bandwidth of interest.
The heart of the frontend is the HFET amplifier (for
frequencies up to 100 GHz) or SIS mixer (for all other
bands). In order to achieve the required low-noise CONCLUSION
performance, these need to be cooled to about 4 K. ALMA is the biggest radio astronomy instrument
Each frequency band has two (dual polarisation) ever realised. Its size, instrumentation and site make it
receivers in one "cartridge", complete with the cooled an instrument of unparalleled performance capability.
parts of the Local Oscillator system and IF-amplifier. It will be a true "World Telescope" in which institutes
from North America, Europe and likely Japan will
These cartridges, coming from different institutes, are have joined forces. The technical, managerial,
inserted in one large cryostat; a cylinder of 1 m logistical and cultural challenges are great. To
diameter and a height of 0.7m. Three stages of low overcome those is the goal of a large group of
temperature (50, 10 and 4 K) are provided to the dedicated astronomers and engineers, as well as the
cartridges for the different devices. The rendering funding agencies. The project is in the development
below gives an impression of the cryostat with phase and construction is to begin in 2002.
cartridges inserted and the "optics" (horns and Completion is scheduled in 2011, with partial science
mirrors) on the top. operation to commence in 2007.