19 Mar 2007
An infrared, 67 Mpixel, 3 tonne camera has been successfully installed on the VISTA telescope in Chile. Martin Caldwell from the UK's Rutherford Appleton Laboratory explains how the instrument pushed the limits of large-scale optics.
One of the world's largest cameras to be built for astronomy was transported to the European Southern Observatory (ESO) in Chile in February. It will be the eyes of VISTA (Visible and Infrared Survey Tele-scope for Astronomy) – a new 4 m-class telescope that will survey the sky in the southern hemisphere in the scientifically important spectral region of 1–2.5 µm.
The telescope's camera is almost 3 m long, uses large optics (i.e. its entrance window is 1 m in diameter) and weighs 3 tonnes. These features make it a real colossus among optical cameras, and its demanding performance requirements (see tinted box) have pushed large-optics technology to the limit.
The design required the finished instrument to be portable – this means that it can be lifted on and off the telescope by a crane as a single assembly. In total, the camera has taken four years to design and build, and uses components sourced from sup-pliers across Europe and the US.
On the move
The camera had to be built on a dedicated handling stand. The stand and the camera's mounting had to be made strong enough to allow for the earthquake-prone environment of Chile. This feature was also useful when it came to transporting the camera. It was just possible to make the camera and its stand fit within the largest cargo container allowed on a Boeing747.
As with other large optics, such as mirrors for astronomical observatories at remote sites, this journey to Chile was particularly risky for the fragile and high-value optics of the camera, especially as the last part of the trip was along a very rough dirt road in the Atacama desert.
Because of such concerns, many previous instruments have been shipped in several parts, but this has other disadvantages. In this case, the bold decision was made to ship the instrument complete. This was made possible partly thanks to the built-in strength of the earthquake-proof assembly and partly by following the approach of the space industry for shipping satellite optical pay loads of a similar size to launch sites.
The camera arrived safely at the ESO at the beginning of February and members of the team have spent a month installing it and carrying out initial tests. This included the delicate operation of lifting the portable camera off its stand by crane and raising it into the observatory dome to finally unite it with its telescope.
Various other tests of VISTA's systems and software will follow and the camera's first use on the sky should start this autumn. There are plans to take test images of so-called guide stars to set up VISTA's active optics system and verify its imaging performance.
After that, the VISTA science programme will use the wide-angle, high-sensitivity properties of the camera and its optics to make new surveys for infrared (IR) objects across the sky of the southern hemisphere. The aim is to look deeper into the universe and yet be able to complete the searches faster than ever before.
Pushing the limits of large optics
The camera uses the world's largest focal-plane array of detectors measuring 0.3 m in diameter and incorporates wavefront sensing detectors for the active optics control of the telescope. The term active optics refers to the compensation of errors such as focus, alignment and aberrations, which are corrected by adjusting some components on the actuators. The two telescope mirrors are made active in this way, using signals derived from the camera wavefronts.
The focal plane is so large because VISTA has a relatively wide-angle field of view, imaging an area of sky that is three times the diameter of the Moon. The cryogenic detector array comprises 16 IR-detector modules from Raytheon, US, giving 67 Mpixel in total. The wavefront-sensing system was built at the University of Durham, UK, using charge-coupled devices from e2v.
Packing such a large format of detectors into one camera produces multiplexing challenges for supplying power, clocking of read-outs and extracting signals at low noise. The camera has a significant amount of electronics on board to overcome these problems, including five large racks mounted around its outer surface. In sustained operation, wide-field images can be produced every 10 s leading to data rates of up to 1 Tb per night of observing.
A giant filter wheel that is 1.4 m in diameter sits just above the focal plane (see diagram). The wheel needs to be large as it has to house up to seven arrays of spectral filters. Each of these arrays covers one waveband and is approximately 0.3 m in size to fully cover the detector array. The science wavebands currently installed are centred around 1.0, 1.2, 1.6 and 2.15 µm, with bandwidths of 0.1–0.3 µm.
Developed by the UK Astronomy Technology Centre (UKATC), the challenge with such a large, moving component was to create a cryogenic mechanism and motor to deal with the large inertia involved when swapping rapidly between different filter sets. Change-over times are on the order of 10 s.
The main camera optics comprises a group of three lenses, each 0.6 m in diameter. The lenses correct aberrations and allow a normally narrow-field two-mirror telescope to be used over a wide field of view. The lenses are the maximum size that can be manufactured from IR-transmitting silica. They are fabricated by Sagem in France and built into a cryogenic lens-barrel by UKATC.
A significant challenge was making the lens mounts stiff enough to limit the flexure effect when the camera orientation varies as the telescope tips around the sky, and yet compliant enough to cope with the large thermal strains that occur during cool-down and warm-up. Both effects are exacerbated as lens size increases due to the low conductivity of silica, and their control is critical to the quality of the final image.
Above the lenses is a large cryogenic baffle tube that blocks ambient thermal emission from the observatory and atmosphere, which would otherwise swamp astronom-ical sources emitting in the 2 µm spectral region. Because the system is wide-field and relatively fast (∼F/3), it is not possible to use the conventional approach where an internal cold pupil stop would fully block this background emission. Instead, blocking is achieved by making the cold baffle tube of sufficient length to prevent direct views of the ambient surround and by making its internal surfaces sufficiently absorbing.
This creates another problem. The baffle must be reflective in the thermal wavelength region of approximately 10 µm, such that heat radiated by the window to the baffle is efficiently returned to the window to prevent it from becoming too cold and misting up. The answer is to use specially designed reflective baffle vanes with a dichroic coating that absorbs in the 2 µm region, but reflects in the longer thermal wavelength region.
Under test, this solution, which includes a coating developed by Reynard Corporation, US, offered sufficient stray-light blocking as well as the correct thermal balance between baffle and window. The required baffle length makes the camera's cryostat relatively long and leads to its front window having an aperture of approximately 1 m.
Such a large vacuum window, fabricated from 8 cm-thick IR-transmitting silica for strength, and to meet stringent optical quality requirements, could only be made with the collaboration of three manufacturers: Heraeus, Germany, which produced a large boule of homogenous Infrasil glass; Corning, US, which flowed-out the glass to the required size; and Sagem, France, which performed the final polishing using ion-beam techniques to remove the residual inhomogeneity effects and arrive at a transmitted wavefront error of approximately 100 nm.
One final challenge was the optomechanical design and producing a structure to hold all of the cryogenic items (detectors, filter wheel, lenses and baffle, totalling 0.7 tonnes) on a thermally isolating cradle in a cryostat vessel. This structure needed to be sufficiently stiff to limit flexing effects and yet compliant to cope with thermal cycling, while allowing significant cooling power to be applied and last, but not least, having the ability to survive earthquakes. Particularly challenging were the issues of controlling the detectors and the optics focus positions to within less than 10 µm planarity across the large focal plane and making this stable enough with respect to the telescope.
• This article originally appeared in the March 2007 issue of Optics & Laser Europe magazine.