Reprinted from the Fall/Winter 1992 Research Review Magazine

Of all the tools and instruments of science, there is none so grand or enduring as the telescope. Looking up at the heavens, a telescope brings us images of the past that say much about the present and provide clues about the future. Some telescopes "see" radio waves, some see x-rays, and some see the heat given off by celestial objects; but the bread and butter tools of modern astronomy remain the optical telescopes, those that see visible light--the principal radiation of the stars.

This past spring, construction was completed on the largest optical telescope in the world, the W. M. Keck Telescope on the island of Hawaii. With a light-gathering mirror that measures 10 meters (400 inches) in diameter, the Keck Telescope will essentially double the observation range of any existing telescope, bringing into view objects more than 10 billion light years away (one light year is equal to about 6 trillion miles).

Built and operated by the California Association for Research in Astronomy (CARA), a partnership of the University of California (UC) and the California Institute of Technology (Caltech), the Keck Telescope represents the first major departure in fundamental optical telescope design since the days of Sir Isaac Newton. This design breakthrough, which has been reproduced in a twin 10-meter telescope adjacent to the first, was born and nourished at LBL.

In 1977, UC formed a five-member committee to come up with a design for a proposed ground-based optical telescope whose reflecting mirror would be double the size of the Hale Telescope at Mount Palomar, the 5-meter (200-inch) behemoth that had been the nation's biggest telescope since 1948. The committee included representatives from each of the UC campuses with major astronomy programs--Berkeley, Santa Cruz, Los Angeles, and San Diego--plus an experimentalist with LBL's Physics Division, whose training and background had been in particle physics, but who had become involved in studies of pulsars and neutron stars. This astrophysicist was Jerry Nelson.

Initially, when the UC committee met, the idea was to design essentially a scaled-up version of the Hale Telescope, a telescope with a monolithic reflecting mirror. This approach presented a number of problems. A reflecting mirror with a 10- meter diameter would require an elaborately complex structural support system to keep it from collapsing under its own enormous weight. Also, the larger a mirror's surface, the thicker it must be in order to withstand gravitational effects that could alter its shape.

As the size swells, the cost of the mirror becomes exorbitant. It was estimated that the cost of a 10-meter telescope, using a single reflecting mirror, could top a billion dollars. Everyone agreed that the prospects for funding a billion-dollar telescope were slim.

Nelson had a better idea.

Instead of a single, gigantic reflecting mirror, Nelson proposed constructing a parabolic or bowl-shaped reflecting surface out of many thin mirror segments. He argued that if the technology of a segmented mirror could be mastered, there would be no inherent limit to the size of a reflecting surface.

Said Nelson in an interview at the time, "The Hale Telescope was very innovative for its day, but in terms of advancing the state of the art--or at least pushing the available technology to its limits--it's been downhill ever since for optical telescopes. It is time for a forward step, not just making improvements in an old design."

The UC committee continued to favor the monolithic mirror approach but was intrigued enough to invite Nelson to develop his idea. He spent the next two years working at LBL with a number of researchers, most prominently, physicist Terry Mast and engineer George Gabor. In 1979, he presented their segmented mirror approach to the committee, and it was chosen over several other concepts.

Specifically, the design that Nelson and his colleagues settled on called for a mosaic of 36 hexagonal mirror segments arranged in the form of a honeycomb and kept in perfect alignment by a computer-operated active control system. Each mirror segment would be 1.8 meters wide, 7.5 centimeters (about 3 inches) thick, and weigh about half a ton. Taken as a whole, this segmented reflecting mirror was expected to be about the same weight as the Hale Telescope's mirror even though it would be quadruple the size in surface area.

THE UNIQUE MOSAIC DESIGN OF THE KECK PRIMARY MIRROR CONSISTS OF 36 HEXAGONAL SEGMENTS.  THESE SEGMENTS ARE ALIGNED TWICE A SECOND TO AN ACCURACY OF FOUR NANOMETERS -- OR 1,000 TIMES THINNER THAN A HUMAN HAIR -- THUS FORMING A VIRTUALLY CONTINUOUS PIECE OF GLASS. Photo courtesy of the W.M. Keck Observatory

Nelson and his LBL colleagues faced two major hurdles: designing a means of keeping the 36 segments perfectly aligned, and finding a way of polishing the segments so that together they would function as a single, giant, parabolic surface.

The surface of the reflecting mirror of any large telescope is continually subjected to stress from such factors as atmospheric temperature changes, shifting winds, or changes in the mirror's position as it scans the sky. Nelson knew that such stresses would readily disturb the alignment of his proposed mirror segments, so developing a system that could compensate for these stresses was essential.

The alignment system Nelson, Mast, and Gabor finally devised consisted of 168 electronic sensors mounted on the edges of the hexagonal mirror segments, and 108 individual, motor-driven adjusting mechanisms, called "position actuators," which are connected to the back of the segments at the rate of three actuators per segment. The sensors on each segment continually compare the height difference between the segment and its neighbors. If a segment moves with respect to its neighbors even as little as a millionth of an inch--a thousand times thinner than a human hair--this information is instantly relayed to a computer. The computer collects data from all 168 sensors, calculates what adjustments are necessary to bring all of the segments back into alignment, then directs the actuators to make these adjustments. By the time one batch of correcting movements has been completed, a new reading has been taken and the process begins again. Alignment of the Keck's entire array of mirror segments occurs twice every second.

In addition to this "active" mirror support system, a team of LBL researchers designed a "passive" support system. The passive support consists of a thin, stainless-steel "flex disk" attached to the back of each segment and three "whiffletrees" with articulating arms that "float" the segment at 36 evenly distributed points, as if there were no gravity. Working in tandem, the flex disk and whiffletrees resist side-to-side motions but permit the up and down and tilting movements needed to align the segments.

Nelson credits much of the solution to the mirror-polishing problem to Jacob Lubliner, a professor of engineering with UC Berkeley. Adapting a technique first used by the German telescope designer Bernhard Schmidt to fashion a combination reflection-refraction mirror into a huge camera, Nelson, working with Mast and Lubliner, developed a technique called "stressed mirror polishing."

In this technique, forces and torques are selectively applied to the edges of a mirror blank in order to warp the blank to a desired degree of distortion. A simple sphere is then polished into the distorted blank, and the forces and torques are removed. If the correct forces were applied, the spherical surface will elastically relax into the desired shape.

Explained Nelson, "Polishing a segment of a parabolic curve is practically impossible, the mirror segment would have to be the shape of a potato chip. But polishing a segment of a sphere is quite routine."

Knowing the final shape of each mirror segment they wanted, it was a simple matter for Nelson and his colleagues to determine the amount of distortion they needed. Solving the equations governing elasticity for the glassy ceramic material they used as a mirror blank yielded the forces needed to achieve this distortion.

Once a mirror blank was polished, it could be cut into a hexagon. Nelson and his colleagues later learned that the cutting operation slightly distorted a mirror's parabolic pattern. However, the attachment of a harness, consisting of 30 metal leaf springs, to the back of the mirror could impose forces on the mirror that would bend it back into the required shape.

In 1984, a technical demonstration of the alignment and control system was held at LBL using a full-size mirror segment and a prototype sensor and actuator. The demonstration proved that the system worked.

During this period, there was also intense activity on the structural design of the telescope, led by UC Berkeley's Steve Medwadowski, after a preliminary design effort led by Lubliner. Techniques for moving the telescope and sensing its position were pursued by Jack Osborne of UC's Lick Observatory. Bob Weitzmann of Berkeley's Space Sciences Laboratory made fundamental contributions to the passive support system and its attachment to the telescope structure. After an intensive review of all of this work, astronomers from California Institute of Technology agreed to join UC in the project. The search for funding began in earnest.

The donor that was found was the W.M. Keck Foundation of Los Angeles. Named for William M. Keck, the founder of the Superior Oil Company, this foundation is among the nation's largest charitable organizations. At the beginning of January 1985, the foundation announced it would grant $70 million to Caltech to build the telescope. Caltech and UC officially formed CARA. Under the agreement, Caltech would provide the money to build the telescope (approximately $94.5 million) and UC would provide the money to operate it for roughly 25 years.

On September 12, 1985, ground was broken to begin construction on a cinder cone atop an inactive volcano called Mauna Kea, widely acknowledged as the best site on the face of the Earth for a telescope. At an elevation of 13,796 feet, the peak of Mauna Kea is above 40 percent of the atmosphere. The thin air up there is dry, usually cloudless, and far removed from the polluting chemicals and urban lights that hamper other observatories. Mauna Kea is also surrounded by a thermal blanket, the Pacific Ocean, which greatly reduces the air turbulence that can disrupt visible light observations.

When CARA's construction of the Keck Telescope began, Nelson was named the project scientist. Named as project manager was Jerry Smith of the Jet Propulsion Laboratory, who had previously managed the construction of NASA's Infrared Telescope Facility on Mauna Kea and who also oversaw the international collaborative effort to develop the Infrared Astronomical Satellite.

LBL's Engineering Division continued to play a vital role in the project. Taking over project manager duties for LBL was Engineering's Andy DuBois. Dick Jared was placed in charge of work on the active support system, and Bob Fulton was named to lead the passive support system work. Steve Lundgren and Bob Minor also played prominent roles. (Jared and Minor will be heading up LBL's contributions to the construction of the twin telescope, called Keck II. LBL will be receiving some $750,000 over the next four years to serve in a consulting role in the fabrication of the Keck II active support system.)

By the summer of 1987, groundwork for the Keck telescope had been completed, the concrete foundation had been poured, and construction had begun on the dome. Experience has taught astronomers that even small sources of heat within a dome can distort or seriously degrade the images captured by a telescope's primary mirror. Consequently, CARA went to great lengths to remove from the Keck's dome all sources of heat. The dome's interior walls are insulated and covered with a paint chosen for thermal properties that minimize internal heating during the day. During the day, the interior is air-conditioned, and huge fans can replace all of the air inside the dome every five minutes if necessary. This keeps the indoor and outdoor temperatures nearly identical and ensures that the quality of the mirror's image is limited only by the natural properties of the atmosphere. The thermal design of the dome was studied and optimized by Bill Carroll of LBL's Energy and Environment Division.

A significant savings in construction costs and materials was achieved because the Keck's dome could be made significantly smaller than the dome of the Hale Telescope. The dome's compact size--it measures 30 meters high and 36 meters wide--was possible because of the extremely short focus of the Keck's segmented mirror and because the Keck uses what is called an alt-azimuth mounting rather than a conventional and considerably bulkier equatorial mounting.

While the dome was being erected, the Keck's tube structure was being assembled in Spain and tested for flexure. This eight- story structure was designed to provide maximum strength and stiffness with a minimum amount of steel. Thanks to its unique design, and to the fact that the mirror segments are so much thinner than a 10-meter monolithic mirror would have to be, the overall weight of the Keck Telescope is 298 tons, about half that of the Hale Telescope.

In 1989, workers began installing the Keck telescope's steel tube structure inside the dome. The space-frame mirror cell that holds the mirror, which consists of more than 1100 individual working pieces, was added next; and on October 12, 1990, the first mirror segment was put into place. Two months later, on December 4, with nine of its mirror segments in place, the telescope captured its first light.

CARA scientists released a dramatic photo of NGC 1232, a pinwheel galaxy some 65 million light years away. Even with only a fourth of its primary mirror installed, the Keck telescope already matched the light-collecting power of the Hale Telescope.

Said an exuberant Nelson during the press conference at which the photo was released, "We are jumping up and down over the sharpness of the image."

At the same conference, Lick Observatory director Robert Kraft said, "The image is pretty spectacular. At this stage of the game, I don't know of any telescope in modern times that worked properly right off the bat like this."

The W.M. Keck Observatory was officially dedicated on November 7, 1991, with a ceremony that included in addition to the usual speeches, a chanter and priest who offered traditional Hawaiian blessings of the telescope and the people who will work with it, and a traditional Hawaiian ground blessing of the site for the Keck II. On April 14, 1992, the final mirror segment was lowered into the cell, and the telescope was complete.

"I'm ecstatic," was Nelson's official comment.

New telescopes generally require a year or more of shakedown time to become fully operational. During this year, Nelson, working with Smith and other CARA members, will be testing the telescope and fine-tuning its optics to obtain the best images possible.

One of the key tests that has already been passed was whether or not the warping harnesses would work as they were supposed to and correct the shape of each mirror segment to within one-millionth of an inch.

"With these corrections, we hope to get a resolution of 1/2 second of arc (astronomers measure resolution in angles, one arc second is 1/3600ths of a degree) or better when the atmosphere is calm," said Nelson. "This is roughly equivalent to being able to distinguish a car's headlights as two separate objects from a distance of 500 miles, and is about twice as good as the Keck's first images."

To capture its images, the Keck will use either a Cassegrain or a Nasmyth focus. In the Cassegrain focus, light from an object being viewed will be gathered by the mosaic of mirror segments and reflected back toward the prime focal point. Before reaching the prime focal point, however, the light will be reflected once again by a small convex secondary mirror, held in a steel frame about 15 meters above the primary mirror, back through an opening in the center of the primary mirror. In the Nasmyth focus, an oblong tertiary mirror is positioned in this opening and it reflects the light from the secondary mirror to foci located on either side of the telescope.

There are two secondary mirrors that can be used on the Keck telescope--one for optical studies and one for infrared studies. One of the first telescopes designed for both visible light and infrared viewing, the Keck's infrared capabilities are expected to be far superior to those of any other telescope now in operation. It will make infrared measurements 40 times faster-- meaning it can see much fainter sources of radiation--and produce infrared photos three times sharper than any telescope before.

The dry air above Mauna Kea is a further boost to the Keck Telescope's infrared capabilities. These capabilities will enable the Keck telescope to penetrate areas of space too clouded by dust and gas for clear, visible-light sightings. Infrared viewing is particularly useful for finding protostars--stars still in the formation process--that radiate a great deal of heat but are enshrouded in an opaque cocoon of star stuff that makes them appear as smudges to an optical telescope.

Whether the telescope is operating as an optical or an infrared instrument, rarely will there be an astronomer peering through an eyepiece at the Keck's images. Instead, images will be recorded by computer-controlled cameras. This is because, to the human eye, a dim object will appear dim no matter how long it is stared at, whereas photographic film, or other types of detectors, exposed to a dim light for a long enough time will produce a bright image. Where the telescope looks will also be computer-controlled. A telephone line linked to the computer allows telescope-pointing instructions to be phoned in.

UC and Caltech astronomers will each be allotted 45 percent of the telescope's observation time, and astronomers with the University of Hawaii will get the remaining time. Some time from each group, however, will be granted to astronomers from other institutions. One idea that will be tested is for astronomers to send their observation requests to a permanent staff at the observatory, which will collect the data and send it back to the requesters. Peter Gillingham, the former officer-in-charge of the Anglo-Australian Observatory in New South Wales, was recently named the first operations director of the W. M. Keck Observatory.

With its gigantic primary mirror and its optical and infrared secondary mirrors, the Keck telescope will be a time machine, enabling astronomers to observe the history of our universe as it unfolds. All optical telescopes look back through time, seeing an object the way it appeared when its light or reflected light originated, just as when we look at someone on the other side of a room, we actually see them as they looked a hundred millionth of a second ago. At such short distances, this time lapse is, of course, insignificant, but it becomes very significant when looking at stars.

For example, some 10 billion light years away from Earth are quasi-stellar objects or "quasars," the most luminous objects in the universe, putting out hundreds of times the energy of the entire Milky Way. Many astronomers believe that quasars are the brilliant cores of galaxies as they were at the beginning of creation. In providing unprecedented views of quasars, the Keck telescope will show astronomers how the universe appeared when time began, shortly after the Big Bang that most scientists believe started the transformation of radiation into matter.

The Keck telescope will also provide a look at the past, present, and future of stars, including our own sun, through detailed spectral analyses of hot, young, blue stars, middle-aged yellows, old and dying red giants, and the final death throes of the white dwarfs. In such an analysis, a prism is used to break down the ordinary white light of a star into a spectrum whose rainbow of colors will be slashed with narrow dark or bright lines. These lines are the spectral signatures of a particular atom or molecule, and from them astronomers can determine a star's chemical composition, its distance from Earth, its age, surface temperature, and gravity, how it is moving through space, and even whether or not it is rotating. Such spectral analysis requires collecting a great many photons, and for that there will be none finer than the Keck telescope with its brute aperture.

The brute aperture of the Keck Telescope will also be brought to bear on one of the greatest searches of all astronomy, the search for stars with planetary systems like that of our sun. In 1996, when the Keck II is scheduled to be completed, the twin telescopes will operate as an interferometer. Interferometry is a technique whereby the light from the primary mirrors of two or more telescopes is converged to form a single image. This not only doubles, or more, the light-collecting power, it greatly enhances resolution, from 1/2 an arc second, which is considered the limit for the best telescopes today, to only a few thousandths of an arc second. Operating as an interferometer, the Keck and Keck II are expected to have enough resolution to search for Jupiter-sized planets around the 100 nearest stars.

Before construction of the Keck telescope began, Nelson was asked what answers to cosmological questions he thought his brainchild would provide. In response, he mused that it might be even more exciting to ponder the questions that the Keck might raise: "With this telescope, we might see something completely unexpected out there, and that's the real reason you build a telescope--to see the things you have not anticipated."

Nelson's sentiments echoed those of Galileo who, in describing the wondrous sights he'd viewed with his telescope, said, "I discovered in the heavens many things that had not been seen before our age."

Sidebar: The Evolution of the Telescope

The telescope was invented in 1608 by a Dutch optician named Hans Lippershey. It was introduced to astronomy in 1609 by the great Italian scientist Galileo Galilei, who became the first man to see the craters of the moon, and who went on to discover sunspots, the four large moons of Jupiter, and the rings of Saturn. Galileo's telescope was similar to a pair of opera glasses in that it used an arrangement of glass lenses to magnify objects. This arrangement provided limited magnification--up to 30 times for Galileo--and a narrow field of view; Galileo could see no more than a quarter of the moon's face without repositioning his telescope.

In 1704, Newton announced a new concept in telescope design whereby instead of glass lenses, a curved mirror was used to gather in light and reflect it back to a point of focus. This reflecting mirror acts like a light-collecting bucket: the bigger the bucket, the more light it can collect. The reflector telescope that Newton designed opened the door to magnifying objects millions of times--far beyond what could ever be obtained with a lens.

Over the next two centuries, there were modifications to the method of focusing, but Newton's fundamental principle of using a single curved mirror to gather in light remained the same. The major change that took place was the growth in the size of the reflecting mirror, from the 6-inch mirror used by Newton to the 6-meter (236 inches in diameter) mirror of the Special Astrophysical Observatory in Russia, which opened in 1974.

The idea of a segmented mirror dated back to the 19th century, but experiments with it had been few and small, and many astronomers doubted its viability. It remained for the Keck Telescope to push the technology forward and bring into reality this innovative design.