Shoot for the Stars

TWELVE SUMMERS AGO, UNIVERSITY of Arizona astronomer Roger Angel swung by a Tucson pottery shop to pick up some firebricks for a backyard kiln. Then he purchased some glass ovenware at a nearby hardware store. A few days later, he materialized in a graduate student’s doorway, brandishing a couple of Pyrex custard dishes melted to a misshapen blob. “We can make telescope mirrors out of this!” Angel exclaimed. Thus began a monumental and quixotic effort to reinvent the central light-gathering surface of the telescope, from its initial design to its final polishing.

This month, many years and millions of dollars later, that effort culminated in a spectacular success: the casting of one of the world’s largest telescope mirrors, a single 6.5-m (21-ft.) circle of glass that sometime in 1994 will be hauled by flatbed truck to the top of Arizona’s Mount Hopkins, where it will tilt skyward like a giant Cyclopean eye.

These are heady days in the rarefied world of telescope making. Not since the 1934 casting of Mount Palomar’s 5-m mirror — a record size at the time — has there been more innovation or competition to push the edge of possibility. In the clear air above Hawaii’s Mauna Kea, the Keck I Telescope’s mammoth 10-m mirror, built of 36 separate segments, is nearing final assembly — a 10-month process was completed last week. Four years from now it will be joined by the Keck II, an equally monstrous twin. By then, the European Southern Observatory hopes to have positioned the first of four 8.2-m telescopes atop a high peak in the Chilean Andes. Japanese astronomers and other groups around the world will be constructing telescopes of similar size and daring before the end of the century.

Collectively, this new generation of ground-based instruments will open an extraordinary new window on the cosmos. “What we can look forward to,” says Caltech astronomer Maarten Schmidt, “is the biggest gain in telescope power in the past 50, maybe even 100 years.” It should bring into focus the most distant quasars yet and even planets orbiting other stars.

The intellectual seeds for this technological renaissance were sown more than a decade ago, when Angel and a handful of other pioneers began contemplating the challenge of building more powerful telescopes. Very quickly, they were forced to consider radical new approaches to mirror design. Simply scaling up old models would have been hopelessly expensive and unwieldy. “A large mirror can’t look like a small mirror,” explains Angel, “for pretty much the same reason that an elephant can’t look like a fly. If it did, its legs would collapse under its own weight.”

The central conundrum confronting designers was this: how to make a telescope mirror that could hold its shape against gravitational sag and gusting winds yet retain the capacity to make rapid adjustments to fluctuating temperatures. As mirror size increases, these two requirements begin to dictate different, and quickly contradictory, solutions. Very thick mirrors resist physical deformation extremely well, but because they retain so much heat, they tend to generate shimmering currents in the cold night air that play havoc with astronomers’ observations. Very thin mirrors, on the other hand, have ideal thermal properties but a daunting physical handicap: as the telescope pans across the sky, a thin mirror will bend and wobble as if made of rubber.

Between this Scylla and Charybdis, mirror designers are charting a variety of bold, new courses. By designing the Keck Telescope mirror as a mosaic of small segments, each the size of a dining-room table, astronomer Jerry Nelson of the University of California, Berkeley was able to make his mirrors both rigid and thin. But to provide images of pinprick sharpness, each segment must be kept perfectly aligned with its neighbors, a task handled by an elaborate electronic network.

By contrast, the mirrors designed for the European Southern Observatory consist of a single, vast expanse of glass, thin (17.7 cm) and very flexible. To control wobbling and stabilize the orientation, these mirrors, like giant catcher’s mitts, will be constantly readjusted by 180 computer-activated steel “fingers.” A prototype mirror has already proved its worth. A flaw identical to the one that crippled the Hubble Space Telescope was easily corrected by adjusting the mirror’s shape.

Angel’s approach relies less on intricate control systems and more on vitreous wizardry. The 10-ton mirror he and his colleagues plan to install in Arizona — merely a warm-up for some 8-m versions — boasts a light-collecting surface that is nearly as wide as a house is tall, yet it averages only 2.8 cm thick. What prevents this marvel from fracturing under its own weight is a supporting truss composed of thousands of glass ribs that are cast as part of the mirror’s underlying structure. Arrayed in a striking hexagonal pattern, the ribs form an airy honeycomb that confers on the mirror the structural strength of solid glass at one-fifth the weight. Because the hexagonal cells are hollow, air can be circulated through them to keep the mirror in constant thermal balance.

Although the conceptual design appears straightforward, the casting of a honeycomb mirror requires considerable technical know-how — and time. Angel’s team tackles the job in their hangar-like mirror lab located, improbably enough, under the stands of the University of Arizona football stadium. In the center of the lab is a huge round furnace. To make a mirror, a complex ceramic mold is assembled inside the furnace and filled with glittering chunks of Pyrex-type glass. Once the furnace lid is sealed, the temperature will slowly ratchet up over a period of several days, at times rising no more than 2 degrees C in an hour. At 750 degrees C (1382 degrees F), when the glass is a smooth, shiny lake, the furnace starts to whirl like a merry-go-round — an innovation that automatically spins the glass into the parabolic shape traditionally achieved by grinding. At about 1150 degrees C, the liquid glass oozes into the mold, filling the cells of the honeycomb.

Cooling the mirror is an equally painstaking process that takes many weeks. Reason: if one section of the glass cools faster than another, it will contract more quickly, creating stresses that lead to cracking. When finally unmolded, the mirror will still require months of tedious polishing to remove any imperfections.

Why devote so much time and energy to increasing the size of telescope mirrors? The quest is driven by science. To understand how the universe evolved from the Big Bang to its present form, astronomers strive to capture ever more fleeting flecks of light that emanated from ancient galaxies billions of years ago. A 10-m mirror increases their chances by providing a light-gathering surface that is four times the area of a 5-m mirror. Even bigger gains will be possible if astronomers proceed with plans to link huge telescopes like the Keck I and Keck II together, combining their light- catching power. The laws of physics serendipitously ensure that such telescopic arrays will also provide sharper images — if spatial distortions in the new thinner mirrors can only be held to a minimum.

Of course, that is a big if. All the new mirror designs are pushing the technological frontier, and already some surprisingly nettlesome problems have arisen. “Naturally, the challenges have come in places we least expected them,” says physicist Terry Mast, one of the scientists who is helping build the Keck Telescope. For instance, the laborious procedure developed for polishing the Keck’s 36 mirror segments turned out to warp them. A system of special harnesses has now been developed to bend the segments to the correct curvature. So far, Angel’s mirrors appear to be free of serious problems, though concerns persist that the honeycomb structure could interfere with “seeing” by leaving a subtle quilted pattern on the surface. Far outweighing any potential negatives, Angel believes, is the fact that his mirrors, unlike the Keck and European mirrors, do not require fancy computerized controls to keep them optimally configured. “When we succeed in casting a mirror,” says Angel, “we’ve produced a piece of glass that makes everything else easy.”

Right now, which design will prove best is anyone’s guess. “We’ll know in 50 years,” says Mast. But whatever the ultimate outcome of this ethereal competition, it is clear that Angel’s creative hand will shape telescopes built for many years to come. He and a team of graduate students are among many astronomers racing to devise an “adaptive optics” system that corrects for the turbulence of the earth’s atmosphere. The system affords ground-based instruments the heady illusion of operating in the clairvoyant emptiness of space. Angel, in other words, is on the verge of endowing his telescope mirrors with wings.