“We’re handling glass that weighs 16 tons and is nowhere thicker than an inch,” says mirror maker Roger Angel as he examines a 6.5-meter mirror for the Magellan telescope from the inside out. The central hole where he is perched will allow starlight to pass to optical detectors from a secondary mirror..

Mirrors, however, had seemingly bumped up against the law of diminishing returns. To make them larger, using conventional design, you also had to make them thicker. Beyond a certain size, however, a mirror will spend the entire night dissipating the heat it absorbed the day before. Heat waves on the surface of a mirror distort the very images the mirror was designed to enhance. Furthermore, the weight of such a mirror, more than 100,000 pounds, would make it virtually unmanageable.

In telescope design there is no Bauhaus paradox: More is simply more

“One obvious alternative,” says Angel, his accent betraying his British origins despite more than a quarter century of living and working in the United States, “is to make mirrors wider but keep them relatively thin.” A thin piece of glass solves two fundamental engineering problems, weight and heat, but introduces a third: flexibility. Even minute changes in shape transform fine optical glass into a sort of celestial funhouse mirror.

JIGSAW PUZZLE MIRRORS

The segmented main mirror of the Keck I has proven so reliable NASA wants to use similar technology for the next space telescope.

Three of the giant new telescopes, the two 10-meter Kecks and the Hobby-Eberly, skirted some of the weight problems associated with casting a mirror from a single piece of glass. Designed by Jerry Nelson, an astronomer at the University of California at Santa Cruz, the Keck mirrors are made up of 1.8-meter-wide hexagons, 36 for each, which when assembled form a perfect parabola. This design made it easier to transport the glass, which was cast in Germany and polished in California and Massachusetts, but also introduced a formidable shaping challenge. Since the precise curvature of a hexagon depends on where it lies in the parabola, polishing had to be subtly customized to each piece. Shaping the Kecks’ segmented mirrors was like painting sections of a landscape on individual pieces of plaster and then assembling them in a seamless mural.

The Hobby-Eberly’s mirror, also assembled from pieces, is a full meter larger than the Kecks’. It has a spherical, not a parabolic, curve, making it fundamentally simpler and not as expensive to build. Each of its 91 segments is congruent with every other segment, which makes them notably less costly to polish. Such mirrors sacrifice sharpness, but HET is used primarily for spectroscopy—the chemical analysis of matter according to the light it emits or absorbs. “The design allows us to get a large number of spectra in a short amount of time,” says project scientist Larry Ramsey of Penn State, which, together with the University of Texas, Stanford, and two German universities, owns the telescope. HET’s simplified mounting mechanism—it can be rotated but is fixed at an angle of 55 degrees above the horizon—produced further savings. In all, HET cost $15 million, compared with $93 million for Keck I and $77 million for Keck II. —M. L.

One group of mirror builders has battled flexibility by constructing its reflectors from a ceramic material that is more rigid than glass and supporting them in a cradle that continually adjusts for any deflection. This is the method being used on the Very Large Telescope and the Subaru. Their mirrors, each no more than about 8 inches thick and weighing about 50,000 pounds, are mounted on a thicket of actuators—hydraulic pistons that make continuous adjustments to restore the mirror’s shape when it is bent by the wind or distorted as it is moved from one position to another. A laser monitors the shape, and the necessary adjustments are calculated and controlled by a computer.

The Mirror Lab’s Roger Angel, who had come to Columbia in 1967 to work in physics but then followed his adviser into astronomy, wanted a more elegant solution: a mirror that was thin and rigid. He started his search in the early 1980s with a distinct disadvantage: “I knew nothing about glass,” he admits. But that didn’t deter him. He began his education, where else, but in the garage of his house in Tucson, Arizona, where he took to melting Pyrex custard cups in a homemade oven to see how hot glass behaved. As the project grew over the years, he moved it into a corner of the astronomy department, then to the University of Arizona’s optical shop, then into an abandoned synagogue, and finally, in 1985, into the new mirror lab he had convinced the department to build.

By then Angel, his collaborator Neville Woolf (another transplanted Brit), and graduate student John Hill had developed a design for mirrors thick enough to be stiff, but also lightweight and quick to dissipate their internal heat: a disk, smooth on one side and honeycombed on the other. “It wasn’t an original idea,” says Angel, who credits George Ritchey, designer of the 60-inch and 100-inch telescopes on Mount Wilson at the turn of the century, with having first dreamed up such a mirror. Ritchey, however, never figured out how to make it.

Rather than trying to carve out excess glass—an unimaginably tedious procedure—Angel and Woolf decided to cast their mirrors mostly hollow. They do this by melting glass over an assemblage of hexagonal columns made from a heat-resistant ceramic foam very similar to the material used in space shuttle tiles. After the glass cools, the foam is pulverized with high-pressure jets of water and removed. What’s left is a thin layer of glass supported by an inch-thick honeycomb structure that is three feet deep at the perimeter and 18 inches in the center. Its weight of 16 tons is about one-fifth what it would have weighed had it been built conventionally. When the mirror goes into use, air will be blown across the honeycomb structure so that the glass will maintain ambient temperature. No nighttime cooldown period will be required.

WHAT'S NEXT?

While many astronomers are queuing up for time on the giant telescopes just coming on line, others are drawing plans for even larger telescopes. If these behemoths are built, they will make the twin Keck telescopes, today’s largest, look like opera glasses.

One future option is to extend the linked-telescope approach of the Keck, the Large Binocular Telescope, and the Very Large Telescope. If linking two or four telescopes gives great results, the thinking goes, why not 16? The telescopes would be placed in a circle like pearls on a two-mile-long necklace, each feeding light to a central facility that would combine the light into one incredibly sharp image. Such a telescope would have the light-gathering ability of 50 Hubble telescopes and the resolving power of a single mirror some half-mile across. It could pick out the flag planted on the moon by the Apollo 11 astronauts or resolve an asteroid near Alpha Centauri.

Another radical plan being pushed by European astronomers is for a single mega-telescope. Dubbed the Overwhelmingly Large (OWL) telescope, it would have a primary mirror made of some 2,000 smaller ones—each the size of the Hubble. Placed together, they would create a single reflecting surface more than 100 yards across with a light-gathering capacity ten times greater than all the telescopes ever built. Even its backers admit it would take a decade to grind all the mirrors, but once built, it could pick out brown dwarfs in neighboring galaxies or supernova explosions from 10 billion light-years away.

Astronomer Matt Mountain, director of the new Gemini Telescope in Hawaii, says this future generation of gargantuan earthbound telescopes would make it possible to study individual stars in some of the earliest galaxies or determine the atmospheric gases of distant planets. But before astronomers get carried away, there’s the question of price. Would these telescopes be worth the billion dollars it would take to build them? “At some point, it becomes more economical to put them in space,” Mountain says. “We’re not at that point yet, but it’s fast approaching.” —J. W.

NO ORDINARY OVEN:

The Mirror Lab furnace spins fast enough to create a deep curve in molten glass.

THE LARGE BINOCULAR TELESCOPE'S MIRROR, UPSIDE DOWN.

The glass will be polished twice, first on the bottom, then the top.

 

 

A SKELETAL STRUCTURE FOR THE LBT:

Eventually the dome will house two 8.4-meter mirrors set 20 feet apart.