Scientists at Stanford and across the globe are at work on a project they expect will revolutionize our understanding of the cosmos. Certainly, the way we look at the universe will never be the same. When fully operational in 2016, the $400-million Large Synoptic Survey Telescope will start producing a 10-year-long movie of the heavens, shot from a mountaintop in Chile by the telescope’s digital camera, the biggest and most accurate ever built. This real-time record will help scientists map the mysterious dark matter and energy thought to make up most of the universe. It will track obscure galaxies, witness exploding supernovae, and issue alerts of potentially hazardous asteroids approaching Earth within 60 seconds of detection.
These images, organized by the most powerful processors and software available, will be accessible to all via the Internet. Apart from dramatic implications for physics and astronomy, project leaders say this endeavor will result in more information for amateur astronomers, new ways to spark schoolchildren’s interest in science, and a host of applications yet to be discovered.
Stanford scientists have been at the forefront of the LSST from the start. Tony Tyson, ’62, proposed the project and directs the nonprofit corporation that oversees it, Tucson, Ariz.-based LSST. Physics professor Steve Kahn, director of particle physics and astronomics at the newly renamed SLAC National Accelerator Laboratory, is deputy project director of LSST and the camera’s lead scientist. Project manager for the camera is SLAC’s Kirk Gilmore, an engineering physicist. SLAC is also involved in developing the data management subsystem.
Tyson, now a physics professor at UC-Davis, imagined building a “dark matter telescope” in the mid-’90s when he was conducting experiments in gravity at Bell Laboratories. That interest was sparked when he was an undergraduate, he says, “through wonderful conversations with [the late physics professors] Leonard Schiff and Bill Fairbank.”
Gravity is a key to understanding dark matter. If you throw a ball into the air, Newton’s law of gravity says it will slow down and eventually fall back to the earth. Think of the universe’s Big Bang as hurling millions of balls into the air. Eventually, they should fall back down, or contract, causing a Big Crunch. Instead, the balls (galaxies) are actually speeding up.
This observation has confused scientists for generations, since it suggests that a law antithetical to Earth’s gravity may be governing the universe’s development. Einstein’s theory of relativity allows for an effect like that if there’s an energy field permeating the universe that exerts negative pressure.
“We have a model for cosmology that works spectacularly well, but we don’t know why,” says Kahn. “We had to throw dark energy and matter into the mix to make these things work, without really knowing what they were.”Astronomers’ best guess is that dark matter is a strange, cool kind of particle created during the hot Big Bang that interacts only weakly with the familiar particles of “normal” matter—the stuff that makes up you and me. In the beginning, the universe was quite homogenous, though some areas with slightly more dark matter exerted a slightly stronger gravitational pull. Over billions of years those gravitational forces grouped matter that formed galaxies. Today, we see those galaxies—the effects of dark matter—but not the matter itself.
Tyson had been trying to map dark matter by measuring how light from distant galaxies moves through it. But like a lone cartographer stomping the millions of square miles of American soil in the 16th century, Tyson realized the telescopic field of view was too narrow. “We desperately needed a new telescope/camera/data facility that could map dark matter over the entire sky,” he says.
“Dark matter causes this cosmic mirage that, for example, can make a circular galaxy look elliptical,” Kahn explains. “If we find galaxies close to one another that are being disturbed in the same way, the pattern will tell us where, and how much, dark matter is present.” The end result could be a map of dark matter, which comprises 96 percent of our universe. This, in turn, will make possible direct measurements and modeling of the universe.
“Dark energy is a cosmic clock,” Kahn says. “If we can predict how long it takes for concentrations of dark matter to grow, and compare that to the red shift of galaxies [the speed at which they are moving apart], we’ll know a lot more about the evolution and fate of our universe.
“In the end, we’ll either find that our theories break down at some subtle level, or continue to work. In a way, it doesn’t matter what the results are; no matter what, we’ll have shed new light on how our universe works, and where it’s going.”
At the Astronomy and Astrophysics Decadal Survey in 2000, the scientific community embraced Tyson’s proposal and renamed his “dark matter telescope” the Large Synoptic Survey Telescope, reflecting its broader mission. The LSST public-private nonprofit partnership was formed in 2003; project manager Donald Sweeney estimates its cost at $400 million. Funding sources include private donors and foundations, the National Science Foundation and the Department of Energy. In an uncertain economic climate, Tyson stresses the importance of a recent announcement establishing LSST as a priority among NSF- and DOE-funded high-energy physics projects for the next five to 10 years.
“This project has gained a lot of momentum,” Tyson says. “And most people are getting involved because of their own interest—not too many are on the payroll.” Twenty-six member institutions and more than 200 scientists and engineers have signed on to help resolve the project’s many technological challenges. They have improved the telescope’s mirror design, software and data-storing capacities.
Because many recently built telescopes attempt to see more distant objects by taking very deep, high-resolution images of a very narrow slice of the sky, the telescopic field of view has been shrinking. Like using a magnifying glass to look at an ant, it’s easy to miss the aardvark meandering a few inches away. Even the most advanced traditional large telescopes rarely catch big events in the act, and when they do, it’s just luck.
The LSST aims to take luck out of the equation. With its 9.62-square-degree field of view and 15-second exposures, the telescope will cover the entire visible sky every three nights, allowing a more thorough, bigger-picture study of a wide range of phenomena.
“For every piece of the night sky we’ll have a movie in color of what’s going on,” says Tyson. “It will be an entirely new window on the universe.”
The LSST will also change how astronomers work. On average, they get access to large, advanced telescopes only a few nights per year due to high demand. This makes progress sporadic at best. But the LSST will be a continuous resource for anyone studying astronomy at any level.
One of the most innovative components of the LSST is its mirror configuration. The naked eye can’t see small, distant objects, just as a digital camera can’t distinguish objects that cover too few pixels. The primary job of a telescope’s mirrors is to collect light from distant objects and to clarify the image that the telescope lens then magnifies, or stretches over more pixels. The more light collected, the wider, deeper and clearer the image.
Most major research telescopes use two mirrors; until now, only one other has ever been constructed with three, and they were much smaller than those planned for the LSST. “One needs three very special mirrors plus an unusual camera to get this big a field of view,” Tyson says in an understatement. “This is far from an ordinary telescope.”
Another first: two of the mirrors have come from one piece of glass, a cost-effective measure that makes the job of continuously aligning them much easier. This two-in-one mirror was cast in Tucson in late March and finished cooling in late August. It will take two years to measure and perfectly polish these mirrors. “Carefully forming three large parabolic glass surfaces will allow us to have extreme control,” Tyson says. “We can specify a mathematical form that gets rid of all aberrations and creates a really precise image.”
Advances in large optics fabrication technology make it possible to create deep curvatures and precisely polish unusual shapes in the mirrors, allowing more light to be focused on their surface. This decreases the telescope’s length and weight, reduces instances of image blurring, and provides the wider field of view critical to the LSST’s mission. The telescope’s field of view will be 3,100 times larger than that of the Hubble Space Telescope, or 40 times the area of the moon.
If the mirrors are the telescope’s eyes, its heart is a $100 million digital camera boasting more than 3 billion pixels. This will be the largest, most accurate digital camera in the world. Measuring 3 by 1.6 meters and 2,800 kilograms, it outweighs an SUV. Unlike traditional digital cameras that operate in high light, the LSST will be accurate enough to detect light bouncing off a golf ball on the moon—and to detect 3 billion faint galaxies.
As with the LSST project generally, much of the technology going into the camera was already available but never before put together. That has attracted scientists from many fields.
“I got involved with the LSST because it is so cutting-edge, and the science that will come out of this project is so extraordinary compared to the kinds of surveys that we are doing today,” says Gilmore, who previously built the focal plane for the Keck Observatory camera in Hawaii.This is not the first time a digital camera has been paired with a telescope, but it is the most seamless pairing ever attempted. Gilmore is working with a team of 70 scientists across 15 institutions to develop and build the camera. He is in charge of delivering 10 subsystems, including sophisticated electronics, optics and charged coupled device sensors. “Each team meets once per week, which is a challenge since the electronics are being developed across nine institutions,” Gilmore says.
These days, many disciplines are producing “a nearly insurmountable data avalanche,” Tyson points out. The volume and velocity of LSST-generated data will pose an unprecedented challenge. Consider the numbers. Each image will have more than 3 billion pixels and the telescope will produce 4,320 new images every 24 hours. For the LSST to work properly, all 3 billion pixels must be read out in two seconds. At four bytes per pixel, up to 30 terabytes (20 million megabytes) of data will be generated nightly. One terabyte is the equivalent of 40,000 file cabinets of text, or a feature film stored in digital form. Eventually, 50 petabytes (50,000 terabytes) of digital information will be stored—the equivalent of 2 billion filing cabinets’ worth of text. This will constitute the largest scientific database in the world, a distinction now held by the SLAC National Accelerator Laboratory with 700 terabytes.
In the face of such a data explosion, Tyson has written, “Discoveries will be made via searches for correlations. The role of the experimental scientist increasingly is as inventor of ambitious new searches and new algorithms.” He sees LSST as a “lighthouse project” in that the solutions to its challenges will be applied in other “big data” spheres. The LSST will require more than a thousand of the fastest processors commercially available, a huge data storage capacity, advanced software, and highly integrated electronics to read out 16 parallel data streams while avoiding system overheating. Because there are hundreds of billions of objects in the sky to fill each image, software will be programmed to “literally think about the universe,” Tyson says.
For example, to sort through and classify data requires 5,000 mathematical operations per pixel. Images will be compared and duplicate light subtracted out, revealing only what has changed—exploded supernovae and other phenomena in real time. The detection of astronomical change could shine new light on many of the oldest mysteries of astronomy. Tyson has a 40-page list of scientific studies to which the LSST will contribute data, so as the software analyzes change and movement, it will also transmit to different places for various applications.
Newly inexpensive disks and fiber optic communications will enable the transmission of all images from Chile to the National Center for Supercomputer Applications at the University of Illinois, where they will be processed, stored and disseminated. With the help of Google and others, those pictures will eventually be available through an Internet search engine. The LSST database will link up with other astronomical archives around the world to create an international virtual observatory. New collaborations will be possible with the 500,000-strong amateur astronomer community in the United States, as well as with planetariums and classrooms across the nation.“If we do it right, the LSST will be a boon for making science education fun for thousands of students,” Gilmore says. A fifth-grade class will be able to choose part of the sky to “observe” periodically, searching for change and discovering for themselves new supernovae, asteroids or comets. High-school students will be able to “fly” through a three-dimensional map of our solar neighborhood with tens of thousands of new asteroids and other objects.
The project will run for 10 years, during which time the telescope will look at every place in the sky a thousand times, according to Sweeney. “We should have enough snapshots by then for a successful cosmological study,” Kahn says. And there is a chance that after that, the telescope’s technology will be upgraded and the survey continued. No changes will be made during its 10-year operational life to ensure uniformity.
Once the LSST is in place atop Cerro Pachón in northern Chile—chosen for its elevation (8,900 feet), good weather, clear skies and existing infrastructure—Tyson assumes it will take about a year to work out the kinks. And then on to mapping dark matter in the universe and possibly, he says, a revolution in physics.
Many scientists want to be a part of that revolution. “When I decided to come to SLAC, I was looking for a project that would marry astrophysics and particle physics,” Kahn says. “The LSST is it.”“This is just the tip of the iceberg scientifically,” Gilmore says. “We’ll be able to do many more things once we get started that we haven’t even thought of yet.” For that reason, the team has been careful to develop software and a survey strategy that are open-minded. “We want to make it possible for some clever student 10, 20 or 30 years from now to say: here’s a great idea—I wonder if the LSST can answer this question,” Tyson says. As the field of view widens, so will the possibilities for the very wide world of science.
AMANDA CRAIG is a freelance journalist in San Francisco. Scott Shackelford is a third-year Stanford Law student and a PhD candidate at Cambridge University.
