The Very Large Telescope: Establishment and Early Observations

The Very Large Telescope, or VLT, is an optical and infrared observing facility built by the European Southern Observatory at Paranal in Chile. Its design includes four Unit Telescopes, each with an aperture of 8.2 metres. The key event in 1998 was formal first light for the first Unit Telescope, UT1, rather than the simultaneous approval, completion and start of regular observations for the entire facility.

According to Claus Madsen's institutional history, when the Council considered the construction proposal on 8 December 1987, Denmark's representative supported the scientific aims but did not yet have authority to make a financial commitment. Swiss representative Peter Creola proposed separate votes on the decision to build and on the funding arrangements. The former received unanimous support; the latter received seven votes in favour, with Denmark abstaining. Denmark subsequently joined fully in 1989. Creating a large instrument first required member states to turn shared objectives into workable financial arrangements.

References: [1] [2] [4]

On 4 December 1990, the European Southern Observatory's Council chose Cerro Paranal. The mountain lies in the Atacama Desert, approximately 130 kilometres south of Antofagasta and 12 kilometres east of the Pacific coast. The announcement explained that the choice rested on more than six years of meteorological comparisons, rather than simply altitude or the clarity of a single night.

Observers used identical measuring equipment at Paranal and the existing La Silla site, recording clear nights and atmospheric stability. A smaller seeing value indicates less spreading of stellar images by atmospheric disturbances. The Chilean government also provided 725 square kilometres of surrounding land to protect the observing environment and reduce the effects of lighting and mining. The document explicitly anticipated that ESO would operate two geographically separate sites. The VLT therefore should not be described as a wholesale relocation of La Silla's telescopes.

References: [2]

Technical records by Tarenghi, Gray, Spyromilio and Gilmozzi show that commissioning in early 1998 initially used dummy mirrors and checked pointing and tracking with a small guide telescope. UT1's control software drew on experience with the New Technology Telescope. A separate control system at the headquarters in Garching, Germany, reproduced problems encountered on the mountain and helped identify corrections.

Altitude oscillations detected in March were eventually traced to the velocity controller. After correction, wind-induced movements still had to be addressed. The team therefore brought forward implementation of field stabilization through rapid movements of the secondary mirror. Test results continually changed the order of work: construction was more than assembling finished components. Mechanical, software and optical specialists had to work together to turn a movable structure into a system capable of recording celestial objects steadily.

References: [3]

After the 8.2-metre primary mirror and 1.1-metre secondary mirror were installed in the second half of April 1998, staff progressively adjusted the optical system, taking exposures first with a guide camera and then with a test camera at the Cassegrain focus. Initially the primary mirror lacked a reflective coating. Aluminization was completed on 20 May, and the mirror was reinstalled the next day. Earlier test exposures and the subsequently announced formal first light were therefore not the same moment.

ESO's original announcement of 27 May identified the acquisition of images suitable for astronomical research during the night of 25–26 May as first light. An active-optics system controlled the primary and secondary mirrors, adjusting mirror shape and optical alignment. This did not mean that every adaptive-optics device for compensating atmospheric turbulence was already complete. First light confirmed a milestone for the first telescope, after which scientific instruments still required commissioning.

References: [4]

The set of images released at the time did more than show the appearance of celestial objects: it also tested the instrument. In a ten-minute exposure of the globular cluster Omega Centauri, stellar images remained round with widths of approximately 0.43 arcseconds, reflecting the stability of sustained tracking. The Cloverleaf quasar is a known gravitational-lens system whose four images have a maximum separation of only 1.3 arcseconds. A two-minute exposure nevertheless resolved them, with a recorded resolution scale of about 0.38 arcseconds.

Each value depended on its observing conditions and cannot be treated as performance guaranteed every night. The announcement also included a short exposure of Centaurus A's dust lane and a composite image of the M87 jet taken on the first-light night. The original records received only basic processing such as flat-field correction, preserving structure at pixel level. Exposure times, filters and stellar image scales made these images inspectable technical evidence, rather than publicity photographs judged solely by appearance.

References: [4]

UT1 carried out science verification from 17 August to 1 September 1998. The team's report records targets including the Hubble Deep Field South, lensed quasars and star clusters. Weather and technical problems prevented completion of some plans. Ground-based multicolour images of the deep field were also combined with near-infrared data from the Hubble Space Telescope to compare objects in the same region at different wavelengths.

Gilmozzi, Leibundgut and colleagues observed at Paranal while the Garching team performed rapid processing and quality checks. Daily videoconferences took place at midday in Garching and early morning in Chile. While staff on the mountain rested, colleagues in Europe continued analysing the previous night's data and sent recommendations before the next night. Observation, recording, correction and feedback formed a continuous collaborative process. After verification, the telescope was returned to the commissioning team.

References: [5]

Gilmozzi's operations report dates the start of regular scientific work with UT1 to 1 April 1999. Fog, cold winds and instrument problems affected the first two days, however, and actual observations of the sky did not begin until 3 April. This distinction shows that an announced opening does not mean data can be obtained every night.

Early work alternated between visitor and service observing modes. Visiting researchers helped choose the next scientific strategy, while observatory astronomers operated instruments and provided assistance. Service mode scheduled projects according to plans and actual weather. The team found that scheduling became difficult if every proposal required excellent seeing and moonless nights. Establishing the VLT therefore also meant creating systems for scheduling, user support and data processing, turning the capabilities of large mirrors into sustained scientific output.

References: [6]