As twilight envelops Mount Wilson, a 5,700-foot peak near Los Angeles, Harold McAlister begins his night of stargazing by retracing the footsteps of the late astronomer Edwin Hubble. Night after night during the 1920s, Hubble headed up this same tree-lined path to scan the heavens through the 100-inch Hooker telescope—the most powerful in the world. What he saw was a bizarre universe extending far beyond the Milky Way, composed of multiple galaxies flying away from one another at breakneck speed. That discovery eventually led to the extraordinary theory about the origin of everything, called the Big Bang. Now, some 80 years later, McAlister pauses along the footpath to gaze with reverence at the huge white dome protecting the famous old telescope. "That 100-inch instrument is more important than the space telescope they named after Hubble," he says. "It's the most important telescope of the 20th century." Then the Georgia State University professor puts his head down and moves on. The stars are crisp above the mountain tonight—a good opportunity for him to stare at them with an entirely new kind of machine for scanning the universe. Passing behind the old observatory, he enters a long corrugated-steel building marked Beam Combining Lab and arrives at the nerve center of an optical interferometer, a revolutionary device scattered across the mountaintop and composed of six conventional telescopes, 3,100 feet of light pipes, and 20 computers. It promises to transform Mount Wilson's reputation from that of keeper of a famous old telescope to the new center of cutting-edge astronomy. This is the largest of a half-dozen interferometers under construction around the world. It is called the CHARA (Center for High Angular Resolution Astronomy) Array, and its ability to see into space with incredible detail—50 times finer than any single-mirror telescope ever built—promises to bring the night sky into incredibly sharp focus. For example, CHARA could zoom in on an illuminated object on the moon as small as a man. "If that man were driving a car," McAlister says, "we could distinguish one headlight from another." More important, CHARA can distinguish one star from another. That may seem odd, but most stars viewed through even the largest and newest conventional telescopes look much as they do to the naked eye—tiny dots of light, dimensionless and deceptive. Spectrographic analysis reveals that most of those pinpoints are likely to be two stars—binaries—or even more stars: Castor, in the Gemini constellation, for example, looks like a single star but is actually six balls of fire dancing around one another. Solo performers like our sun are the exception, not the rule. Soon interferometers will help astronomers figure out why stars tend to flock together and how they behave as they age. Eventually, those lessons will come back home, telling us what our sun was like in the past and exposing threats we can expect from it—giant flares, perhaps, or periods of dimming that could trigger an ice age. Interferometers will open up the heavens anew: "We'll make thousands of stellar measurements that have never been done before," McAlister says. Interferometry is also likely to be a boon to planet hunters. If CHARA can detect individual planets around binary stars, as expected, the census of extra-solar planets will grow immensely. The more planets found, the more likely the prospects of finding planets that could support life. Searching for extraterrestrial planets could be the ultimate fulfillment of Hubble's visionary work that began here more than eight decades ago. "The 100-inch telescope allowed us to think that the universe is broad enough and old enough for many other civilizations to have existed out there," says Robert Jastrow, director of the Mount Wilson Institute. "CHARA will restore the glory of Mount Wilson by examining stars closely for signs of ourselves."
The CHARA (Center for High Angular Resolution Astronomy) Array collects starlight from six separate telescopes via an elaborate conduit system. Two vacuum pipes, eight inches in diameter, protrude from W2, one of a western pair of telescopes. The center pipe carries light from W2; the left carries light from W2's more distant twin, W1. Light beams from all the telescopes ultimately end up in the central Beam Combining Lab.
McAlister enters a clean room in the Beam Combining Lab and slips on booties over his shoes. Inside, CHARA associate director Theo ten Brummelaar fusses over a table of delicate optical mirrors where light waves from CHARA's six separate telescopes are combined. Tired-eyed and unshaven, ten Brummelaar has spent months struggling with complicated calibration problems trying to get all six beams of light to meet at the same spot at the same time—the key to making interferometry work. By contrast, the key to making better conventional telescopes is to build wider and wider mirrors. But both conventional and interferometry telescopes operate on a principle that's not exactly intuitive. When it comes to seeing detail, their ability increases as their baseline measurement increases. The baseline is the diameter across the telescope from one edge to the other; as it increases, the telescope's angular resolution increases. The surface area of the mirror is not important to sharpness and detail. Two small mirrors, one at each end of the baseline, would work just as well as a huge mirror that spans the opening. So scientists began to think about placing individual mirrors much farther apart, collecting their light, and combining the separate light waves from each telescope. The idea was popularized in the late 1800s by Nobel Prize-winner and astronomer Albert Michelson. Michelson took a swath of black cloth and cut two small slits in it, so that when he placed it over the 12-inch lens of his telescope, only two slits of glass showed. He pointed his masked telescope at Jupiter's moons. The moons were dimmer with the mask on because less light was collected. But Michelson discovered that only two small samples of light gave the same angular resolution as an entire 12-inch lens. And using his crude instrument, he was able to measure the diameter of the moons. "All that matters for angular resolution is the length of the baseline," says McAlister, glancing at an architectural drawing of CHARA that hangs on the control room wall. The bird's-eye view shows six small telescopes laid out in a Y formation over the mountaintop, each feeding its collected starlight into the Beam Combining Lab via vacuum tubes. What held true for Michelson's two-holed interferometer with a 12-inch baseline, McAlister says, also holds true for CHARA—a giant six-holed interferometer with a 1,080-foot baseline. But as Theo ten Brummelaar is quick to point out, there is a catch—figuring out how to synchronize light waves from six different telescopes. It requires cutting-edge optics, superfast computers, and new engineering invented from scratch.
The dome housing the 100-inch Hooker telescope, which some 80 years ago gave Edwin Hubble the first view of galaxies beyond the Milky Way, is still a commanding presence atop Mount Wilson. The venerable telescope now has an adaptive optics system that corrects for distortions caused by unequal distribution of heat in the atmosphere. The two small domes are, from left, W1 and W2—CHARA's western telescopes—and the flat-roofed building is the Beam Combining Lab.
In a conventional telescope, the curved shape of the mirror ensures that the distance the light travels from a star to the telescope's detector is the same, no matter where it hits the mirror. In Michelson's mask experiment, the curved lens sent light from each hole to the eyepiece along two paths of identical length, so the two beams arrived in sync. With CHARA, the beams of light from the six individual telescopes must travel through a byzantine network of tubes and mirrors that lead to a computerized detector in the control room. "The separate portions of each little wave have to meet up at the detector and recognize each other as twins, as parts of the same wave," says McAlister. "If they don't arrive at exactly the same time, you see nothing." Of course, light beams from telescopes hundreds of feet apart and at different distances from the detector are not predisposed to converge at the same time. Worse, if McAlister sights a star in the western sky, its light will have an ever-so-slightly shorter trip to the westernmost telescope of the six than it will to the one farthest east. There are even more subtle problems to solve, too, such as tiny vibrations that can raise one scope an imperceptible sliver of an inch closer to a star than another scope. Ten Brummelaar's challenge is to anticipate these light-path-length discrepancies and literally stall any light that arrives early. That is accomplished by "delay lines" that move mirrors up to 160 feet along rails to increase or decrease each telescope's light path. The light from each telescope travels to the combining lab through pipelines that have been pumped free of air. At the lab, each light beam hits a set of mirrors and is bumped onto a delay line, where it bounces back and forth between a mirror at one end of a rail and a mirror on a cart. A computer positions the cart at a nanometer-precise distance along the rail to stall the beam so that it is channeled to a detector at exactly the same time as the beams from the other telescopes. The farther the cart is from the mirror on the wall, the longer the delay. "It's absurd that we have to adjust the light to nanometers after it has traveled all that distance from the star," McAlister says, "but we do."
Top: Together, CHARA's six telescopes compose a light-gathering instrument with a maximum aperture, or baseline, equal to the farthest distance between two scopes: 1,080 feet. The Y configuration allows astronomers to vary the aperture for different observations. Bottom: For an interferometer to work, starlight gathered by separate telescopes must hit a detector at the same time. To compensate for the extra distance light travels to telescope 2, light collected by telescope 1 is diverted precisely the same distance on a delay line. Graphics by Matt Zang
When ten Brummelaar is satisfied that all the optical equipment is in proper alignment, he and McAlister turn out the lights and step into a room in an adjoining building filled with folding tables, old office chairs, and racks of computer equipment. Taking a seat next to McAlister in front of two oversize computer monitors, ten Brummelaar taps out some commands on a keyboard. Several hundred yards away, out in the darkening night, telescope bays open. In the Beam Combining Lab, the delay lines and the movable mirrors adjust in the dark to synchronize the starlight from separate telescopes. Tonight the astronomers are using just two of the telescopes, pointing them at large nearby stars whose diameters have already been measured using smaller interferometers. Before they can zoom in on unmeasured stars, McAlister explains, they must calibrate CHARA using stars whose dimensions are known. Ten Brummelaar aims the two telescopes, and a large white star appears, dancing on the left-hand screen. "It's dancing because of the atmosphere, like your eye sees twinkling," ten Brummelaar says. "But the picture is not the data we're after." Instead, he and McAlister are after a complicated measurement of the "fringes," or interference patterns, of light waves from two telescopes that meet synchronously at the detector. They have programmed the system to represent the fringes as a graph, which pops up on the screen next to the dancing star. After a good deal of number crunching—to be done later, during daylight hours—the graph will show how wide that star is. Surprisingly, astronomers using conventional telescopes have been unable to determine even the basic dimensions of the vast majority of stars, much less examine what their surfaces look like. Most of what we know about stars comes from close-up analysis of just one—our sun. And even so, we know very little. Stellar astronomy, McAlister says, "has been like doing sociology while studying only one person, making broad, sweeping conclusions with an N of one. Really, we don't know: Is our sun a weird Jack the Ripper anomaly, or is it a nice, normal, grandmotherly star?" The first task is to measure the diameter of stars in order to gauge their temperatures. "Temperature is the missing link in astronomy," McAlister says. "Temperature tells us what a star looks like on the inside, how it works." Once he determines the diameter of a star using CHARA, McAlister can look up its total energy output (available from conventional telescopes) and derive its temperature. It only takes a few minutes for ten Brummelaar to "get fringes" on a star and measure its diameter. Soon he will be able to zip through the firmament, measuring—each for the first time—a hundred stars a night. "It will revolutionize the field," says Charles Bailyn, chair of astronomy at Yale University. "These are the fundamental measurements that everything else relies on." The next step in understanding stars is to look even closer—to peek at the details hidden within their diameter. When McAlister takes measurements of a star using several pairs of telescopes, he can use the data to create an image of the star surface and see whether other stars have flares and spots as our sun does. "There's no good theoretical explanation for why the sun behaves like that," McAlister says. These magnetic storms on the sun contribute to global warming here on Earth, and his extensive survey should show whether spots and flares are common and constant on other stars, whether they come and go in cycles of, say, a thousand years, or whether our sun is abnormal for having them at all. We already know that our sun is unusual for living alone. Conventional telescopes fitted with spectrographs have determined that as many as two-thirds of stars are binaries. Even though these telescopes can "see" only one pinprick of light, the signature of a double star shows up as a cyclic Doppler shift in a spectrogram. During one-half of the stars' orbit around each other, one star of the pair is moving toward Earth in our line of sight, and its light blue-shifts in a spectrogram. The other star is moving away, and its light red-shifts. Some time later, as the stars circle each other, the first star starts moving away, red-shifted, while the other moves toward us, blue-shifted. "Binary stars have always been called celestial vermin," McAlister jokes. That's because two stars that look like one when viewed through a conventional telescope can throw off other stellar measurements. "But CHARA," McAlister adds with wry pleasure, "is highly sensitive to vermin." He plans a large census of double stars, measuring their mass, diameter, and temperature, as well as distance of separation and the orbital motion of each pair. The data will help theoreticians figure out why most stars form in multiples, and, by contrast, why our sun formed alone. With CHARA, our understanding of stellar evolution will improve dramatically. So will our understanding of planets. The 100 extra-solar planets discovered in recent years are all associated with single stars or widely separated binaries. Conventional planet detection uses the same spectrographic technique as conventional binary finding—a recurring Doppler shift in the light waves—and you can't look for binaries and planets at the same time. The signals get confused. That won't happen with CHARA. What McAlister proposes is to extend his survey of binaries so that he revisits certain double stars every few months, measuring the distance between them each time. When there are no planets in a binary system, McAlister will see two stars orbiting each other smoothly, like a graceful pair of waltzers flawlessly twirling over time. But the presence of a dark planet will complicate that smooth motion like a mischievous monkey around one or both of the dancers' necks. If McAlister sees a binary star pulled in this way by something he can't see, "we'll call a press conference," he says, because they will have found a planet in a close binary system, a revolutionary find. Greg Laughlin, an astronomer at the University of California at Santa Cruz who studies orbital dynamics, says recent calculations suggest "there's lots of room in binary systems where, theoretically, you could fit happy, stable planets." Using computer simulations based on Newton's laws of motion, researchers have found you could have a planet in a binary system orbit both stars, so long as the distance to the stars is at least 31/2 times greater than the distance between them. Or you can have a planet orbit just one star, so long as it orbits at no more than one-third the distance between the two stars. "Just about every stellar system you can imagine is capable of having stable planetary orbits," says Laughlin. "Some may have habitable planet orbits." But these are still pencil and paper possibilities that scientists can investigate when CHARA and other new interferometers become fully operational. "I can't tell you how long I've waited for something like this," says University of California at Berkeley astronomer Geoff Marcy, the current king of planet finding, who has 70 extra-solar-planet finds to his name. Charles Beichman, who as the chief scientist of NASA's Origins Program is charged with finding life in the cosmos, has equally high expectations for planet finding with interferometers: "If we find that binaries commonly have planets, we double the planet population of the universe. With orders of magnitude better resolution, we're now entering the golden age of astronomy."
INSIDE THE BEAM COMBINING LAB:
(A) CHARA director Harold McAlister, left, and site manager Robert Cadman stand amid optical delay lines, where computer-driven carts on 50-yard-long rails are used to equalize the distance light travels from each of the telescopes in the array to a precision of better than one-millionth of an inch.
(B) When a five-inch-diameter beam comes off the delay lines, it is directed through a scope that reduces it to three-quarters of an inch.
(C) A beam combiner brings together the light from separate telescopes. "This is where the magic occurs," says McAlister.
(D) Technical manager Steve Ridgway watches CHARA associate director Theo ten Brummelaar on the computer that controls the operation of each of the six telescopes as well as all the combining lab equipment.
As another night of tuning up CHARA ends, McAlister steps out of the Beam Combining Lab into the cool mountain air. The stars twinkling over the ghostly white dome of Hubble's grand old 100-inch telescope are fading, and the star closest to us begins to brighten the eastern sky. For McAlister, sunrises and sunsets raise a strange thought: "Is this normal?" If binary stars have planets, and there are more binaries than solo stars, perhaps two sunrises a day is normal. McAlister's work is full of visionary thoughts like that, but on Mount Wilson, the golden age of astronomy is unfolding without the sort of fanfare and shocking pronouncements that issued forth from here in the 1920s, when Edwin Hubble gazed into the Hooker and saw stars beyond our own galaxy for the first time. Hubble was a man for his time, full of grandeur and big statements. McAlister, by contrast, is a man of small things, of precision. The age of interferometry is not about seeing farther; it's about seeing more clearly. McAlister spends his nights delaying light waves with mirrors that must be positioned to the millionth of an inch. The golden age of astronomy is in the details.
Vacuum turning boxes in the Beam Combining Lab render incoming beams of starlight piped in from the six telescopes in the Y-shaped CHARA Array into parallel beams that are then reflected out to the optical delay lines.
The CHARA Array is one of several optical interferometers under construction. At least two others, one in Australia and the Navy's prototype, will have even larger working diameters, or baselines, when fully functional.
NASA's Keck Interferometer on Mauna Kea, Hawaii Baseline: 410 feet. The array of six telescopes includes twin 10-meter (33 feet) instruments—the largest single-mirror telescopes in the world—that will help it examine such faint objects in the northern sky as accretion disks around black holes and protoplanetary disks around young stars.
U.S. Navy's Prototype Optical Interferometer at the Lowell Observatory, Arizona Baseline: 1,430 feet. Navy scientists will use the array of four large stationary telescopes and six movable ones to determine positions on the globe and in space to within less than one-half inch.
European Southern Observatory's Very Large Telescope Interferometer (VLTI) on Cerro Paranal, Chile, South America Baseline: 200 feet. With four 8.2-meter (27 feet) telescopes, the VLTI will use a wide field of view to scan for faint objects in the southern sky.
Space Interferometers NASA plans to launch a single satellite with three dual-telescope optical interferometers on board in 2009. Researchers are also studying the feasibility of eventually launching a space-based array of infrared telescopes, together with a beam-combining satellite, to search for extra-solar planets as small as Earth. — W.S.W.
The CHARA Array Web site: www.chara.gsu.edu/CHARA.
The Web site of the Mount Wilson Observatory: www.mtwilson.edu.
The Optical Long Baseline Interferometry newsletter has links to interferometry projects around the world: olbin.jpl.nasa.gov.
A primer on interferometry is available from NASA's Space Interferometry Mission Web site: sim.jpl.nasa.gov/interferometry.