Piecing Together the Puzzle of Large Mirror Telescopes

When Jerry Nelson was born on 15 January 1944, the world's biggest and best telescope was the 100-inch Hooker Telescope atop Mount Wilson. Nearby, the still-incomplete mirror destined for the Hale Telescope on Mount Palomar was waiting out World War II on the Caltech campus in Pasadena, California. The 200-inch Hale Telescope mirror would not see first light until Nelson was five years old, and it would remain the world's premier telescope until 1993, when it was displaced by the first of two 10-meter Keck Telescopes, which Nelson designed.

Nelson was the first child to go to college from Kagel Canyon, an unincorporated area north of Los Angeles. He picked up an interest in machinery from his father, who developed tools for Lockheed. He attended Caltech in Pasadena as an undergraduate, where he majored in physics and worked part time in the machine shop and building a 1.5-meter infrared telescope. After earning a doctorate from University of California at Berkeley, he was doing experiments in high-energy physics and astrophysics at the Lawrence Berkeley National Laboratory in the 1970s when astronomers began talking about building telescopes bigger and better than Hale.

Nelson and his Caltech classmate Terry Mast had spent endless hours thinking about giant telescopes. They thought Hale was about as big as a monolithic mirror could get, and thought the best way to build bigger telescopes was to assemble them from smaller segments. Soviet astronomers were at the time building a six-meter monolithic telescope in the Caucasus Mountains, but that was only a little bigger than Hale. It pioneered the use of computer-controlled altazimuth mounts when it saw first light in 1975, but it never became a discovery machine like the Palomar Observatory.

US astronomers had even bigger ambitions. In the summer of 1975, Leo Goldberg, director of the Kitt Peak National Observatory in Arizona, proposed building a new giant: a 25-meter Next Generation Telescope. Goldberg had great hopes for a new segmented mirror technology called PALANTIR, an acronym that also refers to the "seeing stone" in J. R. R. Tolkien's Lord of the Rings, but the National Science Foundation turned down his proposal as unpromising.

The University of California had its own ambitions, and in 1977 asked a five-member team to develop plans for the university to build a 10-meter telescope. It included Nelson and professors from the Berkeley, Los Angeles, San Diego, and Santa Cruz campuses. Their first thought was to supersize the monolithic Hale telescope design, but when they added up the numbers the total came in over $1 billion. Nelson suggested instead fabricating the big mirror with segments that could be assembled into a single smooth surface, and his idea wound up on a list of alternatives to be studied over the next two years.

Although a segmented mirror would be a more complex structure than a monolith, Nelson argued that fabricating many small mirrors and fitting them together would be much cheaper and easier. He, Mast, and a small group of others designed a mosaic of 36 hexagonal mirrors, each 1.8 meters wide, 3.5 centimeters thick, and weighing about 500 kilograms each that would form a continuous parabolic surface when assembled on a frame. The total mass would be about that of the Hale mirror, but it would collect light from an area four times larger.

Their approach differed from earlier segmented mirror proposals by using new technology to keep all those mirrors in the proper positions as the telescope went about its observations. They designed an active computerized control system, with 168 sensors mounted on mirror edges and 108 actuators to adjust positions. Their system aligned segments twice a second to within four nanometers. A set of passive supports worked in tandem to resist undesired side-to-side motions. To produce the special shapes needed to form a parabolic mirror, they developed stressed mirror polishing, which deformed the mirror blank into a shape that could be polished to a spherical surface, then would flex back into parabolic shape when released. Nelson's group built prototypes to show their ideas worked, and in 1980 the university agreed to spend over a million dollars to develop the concept far enough for them to raise money for the telescope.

Meanwhile, Geoffrey Burbridge had become director of Kitt Peak and revived interest in a 15-meter version of the National New Technology Telescope that seemed more achievable than Goldberg's earlier proposed 25-meter version. Nelson's segmented mirror technology competed for the project with the University of Arizona's multiple-mirror approach, which demonstrated success in 1979 with the 4.5-meter Multiple Mirror Telescope. Nelson's segmented approach won, but once again, the telescope was not funded.

By 1984, however, Nelson's group had built a full-scale mirror segment, including sensors and actuators, and demonstrated successful operation of the alignment and control system at the Berkeley Lab. That success prompted Caltech to join the University of California program and the W. M. Keck Foundation of Los Angeles to grant $70 million toward building a 10-meter telescope on Mauna Kea. Caltech agreed to provide the rest of the $94.5 million construction funding, with University of California to pay for 25 years of operation.

Construction began in 1985 and Nelson was named project scientist. Thanks to the choice of a short focal length and the use of an altazimuth mount, as well as the design of the telescope, it weighs just 298 tons, about half the Hale Telescope, and the dome is smaller. First light for the completed telescope came on 24 November 1993. By then a second Keck telescope was also under construction; its first light came on 23 October 1996.

Nelson took pains to master the basic principles of segmented mirrors, and it paid off. "We were paving new ground, so it was essential that we had a very deep and fundamental understanding of our design," he said. "This mastery of the underlying principles allowed us to efficiently develop the design and the hardware, and when there were surprises, to solve them." The primary mirror acts like a continuous surface, with less than one percent of incoming light slipping through the tiny cracks between segments.

Today, the world's five largest single-aperture telescopes all use segmented mirrors. So do the two biggest ground telescopes in development—the 39.3-meter European Extremely Large Telescope and the Thirty Meter Telescope—and the biggest planned for space, the 18-segment, 6.5-meter James Webb Telescope. Our view of the skies will benefit from Nelson's work for decades to come.

Jeff Hecht is an SPIE Member and freelancer who writes about science and technology.