Issues Magazine

Here Come the Giants

By Helen Sim

“The Hubble” is winding down, but several large land-based and one space-based telescope are poised to be its successors.

Think of astronomy and you probably think of the Hubble Space Telescope. Launched in 1990, it now seems to have been with us forever. Not only has “the Hubble” been incredibly scientifically productive – by 2011 its data had been used in 10,000 scientific papers – but its images have grabbed the public’s imagination. So much so, they’ve become part of the wallpaper of our lives.

The Hubble is a versatile telescope: its four main instruments observe in the near-ultraviolet, visible and near-infrared regions of the spectrum.

But it’s not a big one: its light-catching mirror is only 2.4 metres across. Many telescopes on Earth are much bigger than that.

So what is the secret of the Hubble’s success?

No air. As the Hubble is a space telescope, it is untroubled by the pesky problems caused by the Earth’s atmosphere. One of those problems is light bouncing around in the air, which makes the sky background not quite black. (At infrared wavelengths the atmosphere itself becomes a source of “light” – it glows in the infrared.) Fortunately, users of Earth-bound telescopes have found ways to subtract much of this background light from images.

More difficult is the problem known as astronomical seeing. The atmosphere shimmers and shakes, due both to wind and to heat rising from the ground. Trying to see through this turbulence is like trying to see to the bottom of a pot of boiling water – well, not quite as bad as that, but you get the picture.

Poor “seeing” takes the point-like images of stars and turns them into blobs, placing a limit on the resolution (degree of detail) a telescope can achieve. An arcsecond is 1/3600th of a degree: it’s the apparent size of a dollar coin seen from 5 km away. At optical wavelengths, the Hubble Space Telescope delivers resolution ten times better – about 0.1 arcsecond. (That’s in practice: its theoretical best resolution is half that.) By contrast, a star observed from the best mid-latitude observatories on Earth appears to be 0.5–1 arcseconds in diameter.

(Interestingly, parts of Antarctica have very good “seeing”. A star observed from Dome C, a high plateau in Antarctica, would on average appear to be 0.27 arcsecond in diameter.)

See http://www.astro.virginia.edu/class/oconnell/astr121/im/seeing-movie.gif for a movie that shows how the atmosphere spreads out and bounces around the light from a star.

Wrestling with Air

Correcting for astronomical seeing has been tough, but during the time that Hubble has been wowing the crowd, the engineers who work on ground-based telescopes have been tackling this problem. The answer they’ve come up with – based on declassified military technology – is called “adaptive optics” (AO). Adaptive optics systems track the air’s turbulence and then feed that information to deformable mirrors in the telescope’s optical path, making them change shape hundreds of times each second to compensate for the jittering atmosphere. These adjustments are minute, and couldn’t be seen if you were to just casually look at the mirror.

To track the turbulence, AO systems monitor the light from relatively bright stars called guide stars. These guide stars can be natural ones or artificial “stars” created by shining laser beams 90 km up into the atmosphere, where they stimulate sodium atoms to glow. For more details, see http://outreach.atnf.csiro.au/education/senior/astrophysics/adaptive_opt...

Adaptive optics wipes out the effects of the atmosphere, allowing optical telescopes to reach their full potential. And that potential depends on their size.

Bigger Is Better

Bigger mirrors collect more light. That makes bigger telescopes more sensitive: that is, able to detect fainter objects (provided that you can do a good job of removing the background light).

More importantly, once you’ve wrestled the atmosphere into submission with adaptive optics, you can enjoy resolution – sharpness of detail – limited only by the size of the telescope itself. So with a bigger telescope you can see not only fainter, but sharper. And that’s what every astronomer wants.

How Good Can You Get?

If two sources are very close together on the sky, a telescope “sees” them as one. How far apart do they have to be for the telescope to distinguish them? That minimum (angular) separation depends on two factors – the wavelength of the light being observed, and the diameter of the telescope – and is called the diffraction limit.

The diffraction limit (measured in radians) = 1.22 × wavelength (in cm) ÷ telescope diameter (in cm).

A radian equals 206,265 arcseconds.

Here Come the Giants

Today’s largest optical telescopes have mirrors that are 8–10 metres in diameter. Using adaptive optics they can already make images sharper than those from Hubble. But in the next 10 years we’ll see true giants: extremely large telescopes (ELTs) working at optical and infrared wavelengths that have light-collecting mirrors 20 metres or more in diameter. ELTs are viable thanks to the adaptive optics technology pioneered with the 10-metre-class telescopes. And they’re possible thanks to another lesson learned with the 10-metres: how to make mirrors that are both large and light.

Ideas for ELTs began to bubble away in the 1990s. But now, like the few leading runners at the end of a marathon, just three ELT projects are headed for the finishing line in the near future. The biggest is the European ELT or E-ELT (formerly known as OWL, the Overwhelmingly Large Telescope). This is a 39-metre telescope designed and to be built by the European Southern Observatory, an inter-governmental partnership of 15 states.

The second project is the Thirty Meter Telescope (TMT): it is managed by a consortium comprising Caltech and the University of California, an association of Canadian universities, and national-level bodies from China, Japan and India.

The third project, in which Australia has a share, is the Giant Magellan Telescope (GMT), a 25-metre telescope being undertaken by a consortium of seven institutions from the USA, two from Australia and one from Korea.

The TMT is to be sited in Hawaii, at the established observatory on Mauna Kea, a dormant volcano. The E-ELT and GMT will be located in the Chilean Andes, on Cerro Armazones and Cerro Las Campanas, respectively. The E-ELT and TMT are currently scheduled to start operating in 2022, and the GMT in 2018. All three projects are underway.

Similar Goals, Different Paths

Not surprisingly, these three ELT projects, like super-maxis in an ocean yacht race, have some similarities. Their science programs, for instance, are broadly similar: all want to attack the problems currently tantalising astronomers. What are the planets around other stars really like, and how did they form? What are dark matter and dark energy? How did the first stars and galaxies form? How have galaxies evolved over time? How are giant black holes related to the galaxies they live in?

All the ELTs increase the astronomical discovery space – the “parameter space” opened up by increasing any instrument’s capabilities. In the case of a telescope, that’s sensitivity, resolution and wavelength. History shows that extending the capabilities of an instrument tends to lead to the discovery of new phenomena. So, astronomers have good grounds for expecting that the ELTs will not just tackle existing problems, but also turn up things they haven’t thought of yet.

While the ELT projects have similar goals, their technical approaches differ. The TMT mirror, for instance, will be made up of 492 individual segments, amassed and aligned to work as a single surface 30 metres in diameter. This is the same approach that the current 10-metre Keck telescopes have used with success: TMT can be considered the “child” of Keck.

The E-ELT also has a primary mirror made of segments – 984 of them. Four other mirrors are used to guide the collected light out to instruments that will analyse it; two of these also provide the telescope’s adaptive optics.

The GMT is different. It will have six off-axis mirrors, each 8.4 metres in diameter, ringed around a seventh 8.4-metre mirror like the petals of a daisy: this arrangement, 24.5 metres across, will be equivalent to a single continuous surface 22 metres in diameter (the length of a cricket pitch). These primary mirrors will reflect light to seven smaller secondary mirrors. The shape of each secondary mirror will be modified at hundreds of points, many times per second: this is the telescope’s adaptive optics system. The GMT will offer images that are ten times sharper than those of the Hubble Space Telescope.

A Final Tweak: Phase Compensation

Even after you’ve corrected for the turbulence of the atmosphere, the incoming starlight will hit the individual mirror segments of the ELT, and eventually the imager that that light is fed into, at very slightly different times simply because the paths the light takes to each segment differ slightly in length. The resulting “incoherence” of the light can blur the final image.

In the GMT, the problem will be handled by using a special camera that takes a phase reference from the light of a star, and uses that to work out the time (or phase) compensation to apply to the light being recorded.

The Giant Magellan Telescope

As the GMT is the project in which Australia has a share, let’s look at it in a bit more detail.

Just as the TMT draws on technology used in the existing Keck telescopes, the GMT mirrors draw on the experience of the Steward Observatory Mirror Lab in Tucson, Arizona, in casting mirrors for telescopes such as the Multiple Mirror Telescope, the two 6.5-metre Magellan telescopes, and the Large Binocular Telescope in Arizona. This lab is the only one in the world making such large mirrors.

To make the mirrors stiff but light, using as little glass as possible – a mere 20 tonnes each – they are cast over a honeycomb mould. The mould is placed inside a giant rotating oven. The glass is placed over it, and heated until it melts around the mould, while the spinning of the oven shapes its surface into a natural parabola. This reduces the amount of grinding required to shape the mirror.

After polishing, each of the off-axis mirrors has two subtle curves, like a saddle or a potato chip. They are smooth to within 5% of the wavelength of light – 19 nm, or a thousandth of the thickness of a human hair. That’s so smooth that, if a mirror was the size of Australia, any bumps on it would be less than 2 cm high.

The first of the GMT’s mirrors was cast in November 2005, and is complete. Mirror 2, cast in January 2012, is having its back surface worked on; mirror 3 was cast in August 2013 and is now cooling; and mirror 4 will be cast in late 2014.

Work is also well underway on the GMT site, which is the peak of Las Campanas mountain on the edge of the Andes, 100 km north-east of the coastal town of La Serena. Just as you’d slice the top off a boiled egg, the top of the mountain has been blasted away, making a flat surface about as big as the Melbourne Cricket Ground. The telescope will sit here.

The general Las Campanas site is already home to the two 6.5-metre Magellan telescopes, so it has roads, power and water in place. It is both high (more than 2550 metres), dry and almost completely bare of vegetation – all factors that make for good seeing.

Australia’s involvement in GMT has been supported with Australian government funding from the National Research Infrastructure Strategy’s 2006 program, followed by funding from the Education Investment Fund. Through these two grants, the Australian government has provided $70 million to partner in the design and construction phase of GMT, and $23 million for GMT-related activities in Australia, including the building of new laboratory facilities at the Australian National University (ANU).

Australia’s share of the GMT is about 10%. This involvement buys Australian astronomers observing time on the GMT, but it also enables Australian companies and institutions to win contracts to build instruments for the telescope. The ANU has been chosen to build the one of the first instruments to go on the GMT (the GMT Integral Field Spectrograph), and it has also been contracted to develop the GMT Laser Tomography Adaptive Optics system concept. The Australian Astronomical Observatory (Australia’s national optical observatory) has carried out a feasibility study for another GMT system called MANIFEST for positioning optical fibres.

The Australian GMT Project Office, hosted by the Australian National University, manages Australia’s participation in the GMT project.

“Bye” and “Hi”

In the next 10 years we’ll be seeing generational change among the world’s major telescopes. Having been serviced five times, the Hubble Space Telescope may now keep working for many years, but will probably bow out by 2020 (by which time it will be 30 years old). But it will hand over the baton to its space-based successor, the James Webb Space Telescope (working at infrared wavelengths) and the ground-based ELTs. And they, we expect, will change astronomy as profoundly as the Hubble has.