More Light!

BY THE END OF THE 1960S, ASTRONOMERS WERE FACED WITH increasing blindness. The largest telescope in the world was the 200-inch Hale Telescope on Mount Palomar in California. Since its dedication in 1948, suburban sprawl, with the accompanying light pollution, had made observations increasingly difficult. Moreover, with interest in space science and astrophysics growing enormously, demand for telescope time had far outrun the available supply. No attempt had been made to top Palomar, and with good reason: Building it had taken 21 years and required millions of dollars and numerous advances in design and materials science. Still, somethi-ng had to be done.

To be sure, great advances had been made in astronomy during the 1950s and 1960s, but most’of them had come in the radio-frequency, X-ray, ultraviolet, gamma ray, and infrared regions. Optical (visiblelight) astronomy had lagged behind. “Yet the discoveries in nonvisible wavelengths had unveiled strange new phenomena that optical telescopes-could elucidate: pulsars, black holes, active galactic nuclei, many others. To spur new telescope construction, in 1969 the Academy of Sciences formed what became known as Greenstein Committee, after its chairman, Jesse Greenstein, of the California Institute of Technology (Caltech). As Greenstein later wrote, he wanted optical astronomy to regain its importance and stop living on the “borrowed glory” of other types of observation.

Telescopes work by collecting the light emitted from distant objects and concentrating it onto a small area. Most early telescopes used lgpses to collect the light, and many small modern ones still do. But for most work done by astronomers, mirror-based (reflecting) telelong since displaced lens-based (refracting) ones. Obviously, the bigger the mirror, the more light you can collect, and collecting more light is desirable for two reasons. First of all, it lets scientists see fainter objects, and second, it allows readings to be faster, so more work can be done in a given period and thus more users can be accommodated. During the 1960s very sensitive means of recording images electronically, rather than with film, let astronomers make much more efficient use of the light that was collected. Opposing this trend was the ever-growing problem of light pollution, as well as the ever-growing population of astronomers. By 1969 the need for more and bigger telescopes was undeniable, and the Greenstein Committee considered what form those telescopes might take.

To the scientists on the committee, many of them veteran astronomers who had been using Palomar for decades, there was no other way to build a telescope. As Nathaniel Carleton of the Smithsonian Astrophysical Observatory notes, Palomar “was so successful that it became the paradigm. Astronomers were satisfied with it.” Still, the committee considered some more radical alternatives, one of which came from Aden Meinel of the University of Arizona and a group of younger astronomers who thought the old way of building telescopes was no longer valid.

During his presentation Meinel said that he had obtained seven surplus 72-inch mirror blanks from the U.S. Air Force, which had purchased them for a space project that was canceled. He described how the mirrors could be used to build six separate telescopes mounted in an array, and their images could then be combined with additional mirrors into a single one that would be six times as bright. It would be designed to collect data in both the infrared and the visible regions, and it would be the equivalent of a telescope with a single 176-inch mirror, making it the second largest in the world, but without the difficulties involved in fabricating a mirror of that size to the necessary ultraprecise specifications. The total cost for such a telescope would be around five million dollars, a fraction of what another Hale Telescope would cost.

According to one committee member, although Meinel’s presentation was “effective and enthusiastic,” the idea was just too fantastic. On first examination, it did not seem feasible to keep six telescopes aligned with the necessary precision. As Meinel recalls, another member (who was in charge of the National Science Foundation’s astronomy program, which historically provides the bulk of public funding for American astronomy) later told him that “he would not give a dime of NSF’s money for this wild idea.”

Frank Low of the University of Arizona, who had independently conceived the multiple-mirror idea at the same time as Meinel and had worked with him on the presentation, remembers the resistance to their new design: “No one had any imagination. … They kept wanting to rebuild the same old thing.” Or as Carleton has noted, “People had this feeling that Palomar had been built by supernatural beings. Building something larger or different seemed almost sacrilege.”

By the time of its final report in 1972, however, the Greenstein Committee was willing to give the multiple-mirror concept a try. It recommended spending $15 million for “advanced sensors and controls,” $5 million for a new 100-inch telescope, $5 million to test the concept of “an optical telescope array” (the design Meinel had suggested), and $25 million for “construction of a large optical array or another 200-in.-class telescope,” depending on whether the test of the array was successful.

The drive to construct larger and grander astronomical telescopes has flourished for almost 400 years, ever since Galileo built his first one, with a lens less than 2 inches in diameter, in 1610. After the completion in 1888 of the first permanent mountaintop telescope, the 36-inch Lick Observatory refractor, on Mount Hamilton in California, the race accelerated. In 1897 Lick was topped by the 40-inch Yerkes telescope in Lake Geneva, Wisconsin, which is still the largest refracting telescope in the world. Next came the 60-inch reflector telescope at Mount Wilson, California, in 1908, which was followed by Mount Wilson’s 100-inch Hooker reflector in 1917.

A decade later, the astronomer George Ellery Hale and Caltech began a project to build what for almost 30 years would remain the greatest of all, the 200-inch telescope that bears Hale’s name. Making it happen took two decades and all the ingenuity that several different industries could provide.

Traditional glass-casting methods reached their limit when faced with the task of making a single huge mirror 17 feet across and weighing 16 tons after polishing. Ordinary glass could not be used because its expansion and contraction when temperatures changed would be unacceptable. Simply pouring a glass mirror of such a size without intolerable imperfections was considered impossible. In addition, the reflective surface needed a base material that was lightweight and extremely rigid, stable against temperature changes, and strong enough to keep from breaking under its own weight as it was aimed at various parts of the sky.

A first attempt by Hale to cast a mirror from quartz failed. After spending more than $600,000, all he had was a cracked 60-inch quartz test blank (a blank is a mirror that has been fabricated but not yet polished or otherwise finished). He then switched to Pyrex, which had been developed in 1915 by Corning and is formulated to expand and contract very little when subjected to temperature changes. Between the initial tests and the final casting, Corning developed a new “super-Pyrex” with even less thermal expansion.

Workers had to make repeated test castings over several years to obtain a usable mirror body. Corning engineers learned that the blank’s weight could be reduced while retaining strength and rigidity by making its front side very thin and casting the rest of the body in a honeycomb pattern. Tests also showed that to make sure the liquid Pyrex flowed evenly into all parts of the mold, the mold had to be heated and stored at a controlled temperature in a special chamber.

When the time came to pour the 200-inch mirror body, the first attempt failed when the heat caused some internal sections of the mold to break free. A second pouring, this one successful, still required almost a year of controlled cooling before the blank could be removed from its sealed chamber.

The weight of this mirror blank—more than 21 tons—caused further problems and expense. Moving it from upstate New York to Pasadena took 15 days and required a special train and route. The support system that was inserted in its honeycomb back to keep the front of the mirror true had to be rebuilt twice. Then, during polishing, which took more than 11 years (including a 3-year hiatus for World War II) and reduced the blank’s weight by five tons, tests revealed that when the blank was put in a vertical position, its weight would cause it to warp slightly. Additional springs and supports were installed to counter this distortion. Finally, a thin layer of aluminum was evaporated onto the glass surface in a vacuum and polished.

To hold the mirror in place required a mounting system that would be strong enough to carry the weight but flexible enough to aim the telescope at the tiniest of distant objects and track them as they moved across the sky. When completed, the entire structure of the telescope, including mirror, frame, and mount, weighed more than 500 tons. This mass was so finely balanced that the motor that turned it was rated at one-twelfth of a horsepower.

By 1948, when it was dedicated, the cost of the Hale Telescope had come to $6.5 million, or more than $60 million in today’s dollars. While scientists celebrated its completion and the potential discoveries it would yield, they also recognized that the expense and difficulty of building a bigger telescope would probably be insurmountable.

Only one larger project, the Soviet 6-meter (236-inch) telescope in the northern Caucasus, entered service during the three decades following Palomar’s inauguration, and this telescope, completed in 1976 at an estimated cost of more than a billion dollars, was far from successful. Its huge mirror was so thick and its enclosure so massive that astronomers could rarely get its temperature settled and matched to the ambient nighttime atmosphere. On most nights its shape remained distorted, out of focus, and unusable.

In the late 1960s a few scientists began thinking about a radical new approach (though not an unprecedented one; the Italian astronomer Guido Horn-d’Arturo had started experimenting with “mosaic mirrors” made of hexagonal tiles in the 1930s). Frank Low, who had been studying very faint objects with telescopes in the 60-inch range, was frustrated by the limitations inherent in their size. Instead of trying to fabricate a single gigantic mirror, with all its problems, he thought: Why not combine the light from a collection of smaller ones? By 1969 he had worked out the details, which he presented in July at a one-day conference on the design of large telescopes that was held at the Jet Propulsion Laboratory in Pasadena. The idea appealed to Low’s boss, Gerard Kuiper, who was in charge of the Lunar and Planetary Laboratory at the University of Arizona. Unfortunately, the laboratory didn’t have the funds to start a project of that size.

Meanwhile, Meinel had had a similar idea. He was a consultant to the CIA and the Air Force’s Special Projects Office, where secret spy technology was being developed. He was involved in designing the Manned Orbital Laboratory (MOL), also known as Dorian or KH-10, which planned to use a Gemini capsule to ferry astronauts to and from an orbiting platform, where they would perform both space science and military reconnaissance. But the MOL program was in trouble. The Air Force was developing much cheaper unmanned reconnaissance satellites that could see objects smaller than one foot across. On June 10, 1969, the Defense Department announced that MOL had been canceled.

Meinel was at a meeting at the Pentagon when he learned of this decision. He knew that seven 72-inch lightweight quartz mirror blanks, designed and built by Kodak for MOL, were finished and ready for use, and he feared that “some security officer would have them destroyed.” Meinel immediately went to Harry Davis, the Air Force Department’s deputy undersecretary for research and development, and told him of his idea. Davis was interested but unconvinced. Meinel returned to Tucson and quickly fashioned a small working model in his home machine shop. When he showed it to Davis two weeks later, Davis agreed to transfer the mirror blanks to Meinel.

Meinel arranged to move the blanks from Kodak’s facilities in Rochester, New York, to the Optical Sciences Center at the University of Arizona. A few days later he called Gérard Kuiper and told him about his windfall. Kuiper asked Meinel to come to his office, then called Frank Low and invited him too. When Low arrived, he found Meinel already there. Although the two had met previously, they had never worked together, and neither knew anything about the other’s ideas.

Kuiper said to Low, “Tell Aden here your idea of a multiple mirror telescope.” For the next several hours the three men chewed over the radical possibility. Meinel (who dubbed the idea Project Colt for its “six-shooter geometry”) then went home and worked out the details, which he published in the journal Applied Optics in November 1970.

Despite getting the mirrors free, the university still lacked the cash to build the telescope. Meinel presented his idea to the optical panel of the Greenstein Committee, hoping its support would help grease the fundraising wheels. The committee initially rejected him, partly because (in the recollection of one member) they misunderstood Meinel to say that the Defense Department would fund the project. Moreover, while radio astronomers had long been doing something similar, to use the same technique with the much shorter wavelengths of optical light would require a degree of precision that most of the committee members believed was impossible.

The idea had its supporters, however. Fred Whipple, of the Smithsonian Astrophysical Observatory, in Cambridge, Massachusetts, was intrigued. The Smithsonian had recently built a 60-inch telescope near the top of Mount Hopkins in Arizona (35 miles south of Tucson), leaving the 8,500-foot summit clear in expectation of putting an even bigger one there. Since then the Smithsonian had been exploring several largetelescope concepts, and it was beginning to favor the segmented mirror idea, whereby scores of hexagonal segments, 40 to 60 inches across, would be assembled into an array 360 inches wide.

Meinel’s Multiple Mirror Telescope (MMT) was a plan to build six separate telescopes; it might more precisely have been called the Multiple Telescope Telescope. As early as 1930 the astronomer E. H. Synge had envisioned a multi-element telescope very similar to MMT’s ultimate design. Whipple’s segmented-mirror concept, by contrast, would create a single image with a single mirror instead of combining a pattern of separate ones. (A third approach, in which the images from separate telescopes would be combined electronically, was abandoned because of cost considerations and worries that the accumulated noise from multiple detectors would limit resolution.)

When Whipple saw the Applied Optics article, he put aside his segmented concept and called Meinel. Both men saw the advantages of a partnership between the University of Arizona and the Smithsonian Astrophysical Observatory. The university would provide the mirrors, while Whipple’s links with the Smithsonian Institution would offer a source of government funding independent of the National Science Foundation. The agreement, signed on December 23, 1971, called for the two institutions to build MMT jointly using six of Meinel’s seven 72-inch mirrors.

Work began the following June. Encouraged by an appropriation from the Arizona legislature, the university designed and built the optics, while the Smithsonian focused on constructing the building and mount. Turning the blanks into usable mirrors required some work. They were made of fused quartz, each a sandwich of front and rear silica plates one inch thick with an egg-crate array of silica stiffeners 11 inches thick in the middle.

The blanks were shipped to Corning and heated in a template to more than 2,800 degrees Fahrenheit to give them approximately the correct curvature. They were then polished, coated with aluminum, and assembled, along with six 10¼inch secondary mirrors, to form six 72-inch telescopes. These were mounted together in a single egg-crate frame in a circle 16½ feet in diameter. Each mirror was held in place by a pair of roller chains and further supported by an air bag pressed against the rear.

Now came the critical question: Could the images from these six telescopes be aligned precisely enough to merge them into a single one? To make this possible required the first use of what scientists today call active optics, a system that makes continual adjustments to the positions of one or more mirrors to align their separate images. Today, such systems can even compensate for tiny changes in atmospheric conditions to sharpen a telescope’s sight.

In MMT’s pioneer active-optics system, a seventh mirror, 29 inches in diameter, was inserted in the center of the six-telescope array. It was used not only as a “gunsight” for aiming the telescope but also as a way to calibrate the big mirrors with one another. A. laser beam was bounced off this central mirror and split by a combination of mirrors and periscopes so that a portion of it was aimed at the edge of each 72-inch mirror. From here each beam was bounced to its own silicon detector, which measured the beam’s displacement from the center of the target and passed the information to a computer. The circuitry then corrected each mirror’s position using piezoelectric crystals.

MMT’s originality did not end with its active optics. The mount that held its six telescopes was another innovation that took advantage of the rapidly advancing field of computers. To simplify the task of aiming at the rotating sky, most previous large telescopes had been installed in what is called an equatorial mount. Using a system of yokes and axles, such a mount is permanently aligned with the earth’s axis, so it and the telescope need turn on only one axis during observations, rotating in unison with the stars as they circle the pole.

Instead of an equatorial mount, MMT used an altitudeazimuth mount. Although lighter and simpler to build, this mount requires simultaneous movement in two axes, a much more complicated operation that was generally impossible for such gargantuans until the advent of computers. Before MMT, the Soviet six-meter was the only visible-light telescope that had used such a mount.

In MMT’s altitude-azimuth mount, the frame holding the six telescopes was placed in a 90-ton Y-shaped yoke. Using it, the frame could be tilted up and down from the horizon to the zenith. The yoke, meanwhile, was attached to a pier in the ground that could rotate the whole unit on a system of ball bearings. To keep a star in MMT’s field of view, the telescope’s computer would simultaneously rotate the yoke and adjust the frame’s tilt.

Surrounding and protecting the entire telescope was MMT’s innovative “dome,” which was far from the hemispheric shape that word usually describes. The six-telescope frame required an unusually wide opening, 27 feet across. Yet the peak of Mount Hopkins, even after its top 25 feet were lopped off, was only 65 feet wide. To build a traditional dome-shaped observatory with such a large slit in such a small space would have been difficult at best.

The design of the building fell to Nathaniel Carleton at the Smithsonian. At first he envisioned the dome as merely a roof to protect the telescope. The observation floor would be a circular turntable allowing access to the focal point, attached to the telescope’s yoke just below the horizon. This design, however, meant that to keep the horizon unobstructed, nothing else could be put in the dome, an unacceptable waste of space.

Carleton then wondered if they could instead attach everything in the dome to the telescope, so that as the telescope rotated, so would the building. That way, it would be possible to install rooms on either side of the viewing slit. Moreover, adding the rooms would require more structural crossbeams, which in turn would increase the strength and rigidity of the whole structure, further protecting the telescope from vibration.

He imagined this rotating building as a kind of mailboxshaped structure, with its viewing slit unveiled like the top of a roll-top desk. Because the astronomers assumed that curved walls were necessary to reduce wind turbulence, which could disturb the clear and calm air they needed, Carleton initially designed the building’s walls to bulge outward.

With this design concept in hand, Carleton began looking for an architectural firm. The architect Peter Floyd pointed out that the building would be far cheaper and easier to build if it had as many straight lines as possible. “What you actually want is a barn on a turntable,” Floyd told Carleton.

Carleton then did some careful analysis and discovered that the wind stress around a square building would be little different from that experienced by a round one. Floyd’s firm got the contract, and the “dome” turned into a 450-ton square building that somewhat resembled a battleship’s gun turret. To rotate, the building was mounted on four 3-foot steel wheels 5 inches wide, which were placed at the building’s four corners and ran on a circular steel track. The telescope itself was anchored in bedrock and turned on its own bearings.

To get electricity into a moving building, the wires were fed up through the telescope’s central pier and then draped down the wall of the pier in what was called a maypole. This allowed enough slack so that the building could rotate 270 degrees in either direction without tangling or breaking the wires. A small moat around the building kept out snakes and dust.

When MMT was completed in 1979, its total cost was $8 million, or about $25 million in today’s dollars. The revolutionary giant proved that with daring engineering, bigger telescopes could still be built, and for far less money than with conventional methods. Following this success, the reluctance of astronomers to try new techniques evaporated. Soon there was a rush to design and build a flock of innovative goliaths. Inspired by the MMT, engineers and scientists even took a fresh look at old designs and found ways to make them work better.

The first to follow MMT were the University of California and Caltech, which in the late 1970s began research into what would become the Keck Telescope. Today it is the largest telescope in the world. Located at an elevation of 13,700 feet on the peak of Mauna Kea, in Hawaii, Keck was completed in 1992 and shattered all previous size records with its 36-section segmented mirror 393 inches across. As is true of many other telescopes, temperature and vibrations are so carefully controlled that no one is allowed inside the working enclosure during the telescope’s operation, since body heat would deform the mirrors.

Even more innovative was the Keck Institute’s decision to build an identical telescope a few hundred feet away. Keck II has been in operation since 1996. The institute will soon use electronics to combine the images from the two telescopes in a giant interferometer.

Then there is the 327-inch Subaru Telescope, also on Mauna Kea, which was completed in 1999. Unlike MMT and Keck, Subaru has a single unsegmented mirror more than 50 percent larger in diameter than Hale’s. To build this leviathan, the Japanese made it just 8 inches deep and devised an active optics system with 261 controlling actuators to adjust regions of the mildly flexible mirror to keep its shape true to within 140 angstroms. A similar system is used in the twin 315-inch Gemini telescopes in Hawaii and Chile as well as in the four 323-inch telescopes that compose the Very Large Telescope (Telescopio Muy Grande) in Chile. VLT, like Keck, will be used as a massive interferometer.

Another innovative giant is the Hobby-Eberly Telescope of the McDonald Observatory on Mount Fowlkes in Texas. This Texas-sized telescope, which opened in 1999, is made of 91 segments that combine to create a mirror 362 inches across. It was built for only $13.5 million. The low cost was achieved by having the mirror fixed in place on the ground, able to rotate but not change its angle of elevation. Instead of moving to look at an object, the telescope lets the earth move to bring the object into its field of view. Then, as in the radio telescope at Arecibo, Puerto Rico, a small mirror suspended above the immobile large mirror is moved to bring and keep targets in focus. Although such a scheme limits what can be observed, it saves so much money that it makes the gigantic telescope (which is used almost entirely for spectroscopy) affordable. And since this arrangement makes available 70 percent of what a moving telescope would, the benefits far outweigh the negatives.

Meanwhile, the twin Magellan telescopes in Chile and the dual Large Binocular Telescope, currently being installed on Mount Graham in Arizona, use their own novel technologies. The mirrors—256 inches apiece in Magellan and 330 inches in LBT—are single elements, but unlike in Subaru and Gemini, they are thick enough to hold their shape. To reduce the weight and cost, the designer, Roger Angel, a veteran of the MMT project, went back to the concepts used by Corning to forge the Hale Telescope’s 200-inch mirror and revived them with a few twists.

To save weight, the mirrors have honeycombed cores, but far more so than on Hale. To reduce the time needed to grind and polish each mirror’s front surface, its mold was spun during melting. Imagine the mold as a bowl of soup. When you spin the bowl, the soup’s surface forms a concave shape. In similar fashion, the mirror’s mold is placed inside a furnace and spun at about seven revolutions per minute. When the borosilicate glass (very similar to Pyrex) melts, its top surface naturally curves into the desired parabolic shape. The result is a rigid mirror that is far easier to polish into shape and far lighter than anything conceived by George Hale when he began building his telescope. Angel developed the spin-casting process in his back yard using a homemade kiln and Pyrex custard cups.

Finally, there is MMT itself. Today the original six-mirror telescope that sparked this renaissance in telescope design no longer exists. As a prototype, MMT proved that innovative telescopes could be built. As a prototype, it also had its share of troubles. Though its active optics worked, the system was plagued by problems, not least the tendency of insects, especially moths, to interrupt the laser beams.

Over time the scientists found that they didn’t need the lasers. MMT was stable enough and precise enough for them to align its six small telescopes by eye and hold that alignment long enough to make observations. “We found that we had a better telescope than we had dared dream of,” says J. T. Williams, MMT’s operations manager.

Nonetheless, by the 1990s the decision had been made to replace the six mirrors with a single one 256 inches across. This new telescope was dedicated in May 2000. MMT once again became a test-bed for new telescope ideas, giving Angel his first opportunity to try making hollow single mirrors by spinning them. Following this success, the University of Arizona (along with other institutions) built a twin of MMT at Las Campanas Observatory in Chile. The twins will soon become triplets, whe in late 2002 the university activates another giant telescope at Las Campanas.

Innovation has changed the process of recording data as well. Charge-coupled devices, like those found in digital cameras, are supplanting the use of film. This change, along with continuing advances in computerized control, means that astronomers are almost never on-site when their observations are being made. Indeed, with parameters for an observation programmed by technicians and data collected digitally and transmitted, many astronomers have never even seen the telescopes that they work with.

The development of large telescopes over the last 30 years shows how technological change often follows unexpected paths. MMT was built because single-mirror technology seemed to have reached a dead end. It made many important discoveries, yet its most lasting legacy lay in the impact it had on the profession. The determination of its creators to try something different—indeed, almost zany—convinced telescope makers worldwide to do the same, thereby producing new telescopes that are even bigger and better than MMT. And in an irony that no one could have predicted, the old technology of single-piece mirrors, whose obsolescence was responsible for MMT’s being built, is now making a strong comeback—with the help of active optics, which the builders of MMT pioneered.