Chapter 7

A Picture’s Worth

Images of Distant Worlds

There is something exciting about a picture of a planet circling another star. For most people I’ve spoken with, reading about hundreds of planets discovered through Doppler surveys, transit searches, and microlensing just does not compare with seeing an actual photograph of one. Somehow the photo makes it a “real” world, even if it is just a faint dot next to a bright, overexposed star. These people will be happy to hear that the era of direct imaging is here at last.

For an alien photographer that wants to capture a family portrait of our solar system from tens of light-years away, the challenge would be akin to catching a glimpse of a few fireflies right next to a bright searchlight from a distance of 1,000 kilometers. Seen from afar, the Sun’s brilliance would completely overwhelm the Earth’s feeble glow. Our home planet is billions of times fainter than its parent star in visible light. Even mighty Jupiter, three hundred times heftier than the Earth, would shine only a bit brighter.

Yet there I was, quoted in the Washington Post on January 8, 2002, as saying “It’s technically now possible to directly image a young Jupiter around a nearby young star. We have not directly imaged a young planet yet. . . . But it could very well happen in the next few years.” The quote came from a press conference held the day before at a meeting of the American Astronomical Society in Washington, DC. A skeptic might have dismissed it as an overly optimistic remark from a naive young researcher—and some telescope-time allocation committees seemed to share that view. But two years later, I was willing to stick my neck out again, this time to a reporter from Sky & Telescope magazine: “If nature did not put any Jupiter-mass planets out in 20 AU orbits, we might come up empty-handed. But if these planets exist in any substantial numbers, we should be able to detect them with the current technology.”

There were several reasons for my optimism. For one, young planets are much hotter and brighter than old planets, hence they are easier to detect. We are used to thinking that planets only shine by reflected sunlight. That is mostly the case in our solar system, some 4.5 billion years after its formation. But planets are a lot hotter when young, thus they emit much more energy than they receive from the host star. Gas giants like Jupiter convert gravitational energy into heat as they shrink in size. The pace of contraction is rapid to begin with and slows down over time. During the first few million years of its life, Jupiter would have been about ten thousand times brighter than it is today. Second, the contrast between a star and its planet is not as bad—though still by no means easy to overcome—if you observe at infrared wavelengths. That’s because planets, even young ones, are much cooler than their stars, so the bulk of their emission comes out in the infrared, where the stars are generally fainter than in the visible portion of the spectrum. As a result, a young giant planet would be only about 10,000–100,000 times fainter than its star in the near-infrared. (The exact ratio depends on the mass and age of the star and the planet as well as on the specific wavelength of observation.) It may not sound like much of an improvement, but it is a lot better than grappling with a contrast of several hundred million or more.

Sharper Visions

The third reason is that a technology called adaptive optics was coming into routine use in the early 2000s at several of the world’s largest telescopes, allowing them to “see” almost as clearly as if they were in space. The Earth’s churning atmosphere not only makes the stars twinkle but also spreads their light into fuzzy blobs when viewed through telescopes. Planets next to their stars would be hidden within those blobs. The blurring effect is so strong that images taken with a 10-meter behemoth, like the twin Keck telescopes in Hawaii, are no sharper than those taken with a 10-centimeter amateur instrument, even though the bigger telescopes collect a lot more light and thus can detect much fainter objects. That’s a major reason for launching observatories like Hubble into space. In 1953, Horace Babcock, an astronomer at the Mount Wilson and Palomar Observatories, proposed a way to overcome the effects of atmospheric turbulence without leaving the ground.

The idea is to measure the changing distortions of light waves as they travel through the atmosphere and compensate for them hundreds of times per second by flexing a thin deformable mirror in the light path. The mirror is pushed and pulled from behind by hundreds of tiny motors, called actuators, so that its shape exactly cancels out the effects of roiling air above the telescope, thus making fuzzy pictures sharp again. It is placed along the path that light takes between the main telescope mirror where it is gathered and the detector—camera or spectrograph—where it is recorded. Building such a system requires a number of sophisticated elements: a fast-acting CCD to serve as the sensor of changing wave fronts, computers that can calculate the image distortions quickly, small but effective actuators that can push and pull different parts of a mirror from behind to change its shape, and mirrors thin enough to respond rapidly and robustly.

The military and the aerospace industry built the first adaptive optics systems in the late 1960s and the early 1970s. As you might imagine, the atmosphere blurs images taken from space of objects on Earth, just as it does our images of stars taken from the ground. Thus satellites monitoring the Earth need adaptive optics too. Academic scientists like Freeman Dyson at the Institute for Advanced Study at Princeton and François Roddier at the University of Hawaii also helped develop the technique. Luis Alvarez’s research group at the Lawrence Berkeley Laboratory in California did one of the first astronomy experiments, with a simple de-formable mirror that applied a one-dimensional correction to demonstrate the improvement on a star’s image. Declassification in the early 1990s, following the end of the Cold War, made military technologies available for civilian use, fueling rapid advances in astronomical adaptive-optics systems. One limiting factor for astronomy is the availability of “guide stars”—stars bright enough for their distortions to be measured hundreds of times a second. For the adaptive optics correction to be effective, a guide star has to be right next to the target of observation. Astronomers, of course, would like to observe targets all over the sky, not just those that happen to lie close to bright stars. So they have resorted to making artificial “stars” with laser beams wherever they need in the sky. A common choice is a laser tuned to a yellow color to excite a layer of sodium atoms about 100 kilometers up in the atmosphere. The sodium atoms glow in a small spot, creating a fake star that can be used to measure atmospheric turbulence. A laser is usually not necessary for planet-imaging surveys though, because the target stars themselves are bright enough to serve as guide stars.

Figure 7.1. What appears to be a single star (top) is resolved into a pair (bottom) with adaptive optics. Credit: European Southern Observatory

As several astronomers initiated direct-imaging surveys, the big unknown was whether nature produces planets bigger than Jupiter in sufficiently large numbers and in sufficiently wide orbits for us to resolve one next to a young star in the solar neighborhood. Getting telescope time was a challenge too, given the always-tough competition at premier facilities and some scientists’ inclination to dismiss these searches as “mere fishing expeditions” without a compelling theoretical case for super-Jupiters forming tens or hundreds of AU from their stars. Tentative support for far-out planets did come from observations of rings, gaps, and clumps in dusk disks around adolescent stars. Some theoretical models invoked the gravitational influence of unseen planets in wide orbits to account for these asymmetries. After all, shepherding moons are seen to keep gaps in Saturn’s ring system clear of particles. But the models could only provide a lower limit for the mass of the planet responsible. Those estimates often came in at just a few times the Earth’s mass; such objects would be too small and too faint to image at present. A few theorists also suggested that gravitational close encounters between planets early in a system’s history could kick some of them out to distant orbits or even entirely beyond the star’s gravitational clutches. It is usually the smaller planets that receive the biggest kicks and are hurled out farthest though, so again they would not be detectable.

Also, imaging a planet is not simply a matter of finding a faint dot next to a nearby young star. The dot could easily be another star in the same line of sight that appears faint because it lies much farther away. In fact, over the years, several research teams found candidates that seemed promising at first but turned out to be unrelated background stars upon closer inspection. To be considered a planet, a candidate has to pass two crucial tests. First, its colors and spectra should show that it is too cool to be a star or even a brown dwarf at its age. Second, it should share the same motion through the sky as the star, thus confirming the two are gravitation-ally bound to each other. The case is even stronger if the candidate’s orbital motion around its star can be measured over time. For years, no directly imaged planet candidate met these criteria.

The first credible claim came from a somewhat unexpected corner. On September 10, 2004, Gael Chauvin, then a postdoctoral researcher at the European Southern Observatory (ESO) in Chile, and his colleagues announced a faint companion candidate next to a brown dwarf with the catalog number 2MASSW J1207334-393254 (or 2M1207, for short). The brown dwarf belongs to a nearby group of 8-million-year-old stars known as the TW Hydrae association and is estimated to be only about twenty-five times more massive than Jupiter. Infrared images taken with ESO’s Very Large Telescope using adaptive optics showed the companion candidate is redder and fainter than its primary, and thus likely to be planetary-mass. Chauvin and his collaborators had managed to take a spectrum of the candidate, which revealed absorption features due to water vapor in the object’s atmosphere and confirmed its relatively low temperature. The comparison of the spectra and the brightness at several wavelengths with those of theoretical models gave a mass estimate of about five times that of Jupiter. The separation between the pair was inferred to be about 55 AU, given the brown dwarf’s distance estimate at the time of 230 light-years.¹

The announcement took a cautious tone. The paper reporting the discovery was titled “A giant planet candidate near a young brown dwarf.” Even the ESO press release retained a question mark: “Is this speck of light an exo-planet?” That’s mainly because the researchers had not yet confirmed whether the brown dwarf and its candidate were gravitationally bound to each other. The confirmation came the following year, when a comparison of images taken a year apart showed the two moving together in the sky. “This new set of measurements unambiguously confirms that 2M1207b is a planetary mass companion to the young brown dwarf 2M1207A. The image released last year is thus truly the first image ever taken of a planet outside of our solar system,” Chauvin told the media.

Still, there is some debate about whether to call this companion a planet because it orbits a brown dwarf rather than a star and because the two probably formed like a binary star system in miniature. Here, the companion is only a few times less massive than its host, whereas that ratio for giant planets around stars hovers between a few hundred to a few thousand. As my research group and others have shown, brown dwarfs are born with disks (see chapter 6), but those disks do not contain sufficient dust and gas to make a planet as massive as 2M1207b. Chauvin acknowledged that “given the rather unusual properties of the 2M1207 system, the giant planet most probably did not form like the planets in our solar system. Instead it must have formed the same way our Sun formed, by a one-step gravitational collapse of a cloud of gas and dust.” Later observations by a team consisting of Subhanjoy Mohanty, now at Imperial College in London, myself, and others raised the companion’s mass to eight times that of Jupiter and also found evidence for a dust disk around it, just as a disk surrounds the primary. Our findings bolster the case for a binary-like formation scenario.

First around a Young Sun

After the 2M1207b find, the race to image a planet around a normal star rather than a brown dwarf continued, using powerful telescopes on the ground and in space. Several groups reported brown dwarf companions close to but above the deuterium-burning boundary of about 13 Jupiter masses. Among those searching was David Lafrenière, then a graduate student at the Université de Montreal. He fine-tuned a technique that others had developed for subtracting a star from itself in consecutive images taken with adaptive optics as the sky rotated during the observations. The idea was to remove the star’s glare as much as possible to reveal otherwise unseen companions in its immediate vicinity. His advisor, Rene Doyon, assembled an international team (including me) to compete for time on the 8-meter Gemini telescope in Hawaii to survey eighty-five nearby, youngish stars. The survey formed the core of Lafrenière’s PhD thesis. Unfortunately, he came up empty-handed, but he was able to place a useful upper limit: fewer than 8 percent of Sun-like stars have a companion greater than 5 Jupiter masses orbiting between 30 and 300 AU.

Once he completed his PhD, Lafrenière moved to Toronto as a postdoctoral researcher to work with Marten van Kerkwijk and me on a project we had started with our previous postdoc, Alexis Brandeker, now back in Sweden. Our project targeted nearby star-forming regions in search of stellar and substellar companions. On April 27, 2008, as he analyzed adaptive optics images from the Gemini telescope in Hawaii, Lafrenière noticed faint objects next to two of the stars in the Upper Scorpius association, roughly 500 light-years from Earth. If they were true companions, one would be a low-mass brown dwarf and the other a planetary-mass companion. After some discussion, Lafrenière, van Kerkwijk, and I decided to request director’s discretionary time from Gemini to take pictures of the two targets at two other wavelengths first; if the candidates’ colors turned out to be very red, implying relatively cool temperatures, we would also take their spectra. The follow-up images showed that one was bluish, thus likely a background star, but the other remained a good prospect for being a planet. So we proceeded to get a spectrum over the summer—and it clearly showed strong features due to water vapor and implied a temperature of about 1500 degrees Celsius, much hotter than Jupiter but too cool to be a star at the age of Upper Scorpius. What’s more, the shape of the spectrum also suggested a young object that had not fully contracted yet. From all indications, it appeared to be the real thing!

Figure 7.2. The planetary companion of the young Sun-like star 1RXS J160929.1-210524. Credit: D. Lafrenière, R. Jayawardhana, and M. van Kerkwijk (University of Toronto)/Gemini Observatory

We were excited but also cautious. We discussed intensely for days what else could mimic the observed properties of this object. Could it be a bloated red giant far away that happens to lie in the same sight line? Could it be an old brown dwarf in the foreground, floating between the star and us? We made comparisons with spectra and colors of many other known objects, and with theoretical models. We showed our results to a number of colleagues for their reactions. In the end, we ruled out all other possibilities and were confident that it is indeed a planetary mass companion. We estimated its mass to be about eight times that of Jupiter. It lies roughly 330 times the Earth-Sun distance away from the 5-million-year-old Sun-like star with the boring name 1RXS J160929.1-210524. What we did not have yet was confirmation that the star and the planet candidate are indeed bound together. It would take at least a year to measure their motion through the sky. In light of the competition, we could not afford to delay an announcement for that long. Given our confidence in the companion’s nature from its spectrum and colors and the extremely low likelihood of a chance alignment between it and the star, we decided to go public (just as Chauvin’s team had done with 2M1207b in 2004). We submitted a paper to a journal, posted it on a preprint server to alert the scientific community, and informed the Gemini Observatory. Gemini planned to issue a press release; meanwhile, one journalist found the preprint online and e-mailed us.

“This is the first time we have directly seen a planetary mass object in a likely orbit around a star like our Sun,” Lafrenière told the media. “If we confirm that this object is indeed gravitationally tied to the star, it will be a major step forward.” The existence of a planetary-mass companion so far from its parent star came as a surprise, and poses a challenge to theoretical models of star and planet formation. Thus, I added, “this discovery is yet another reminder of the truly remarkable diversity of worlds out there, and it’s a strong hint that nature may have more than one mechanism for producing planetary mass companions to normal stars.” I was referring to the possibility that this companion, about one hundred times less massive than its star, might have formed through direct gravitational collapse instead of growing in a protoplanetary disk. Some theorists brought up the possibility that it was ejected to a large distance as a result of gravitational encounters with its (unseen) siblings. A handful of researchers remained skeptical that the object is a true companion. Since then, we have made follow-up observations in the spring and summer of 2009 and confirmed that it is indeed moving through the sky with its star. So its existence as a companion is no longer in doubt, but questions about its origin remain.

Siblings Unveiled

In many ways, the fall of 2008 marked the beginning of a new era in extrasolar planet research. Barely two months after our announcement, two other teams reported planet images in the same issue of Science. One group, led by Christian Marois of the Herzberg Institute of Astrophysics in Canada, found three massive planets circling HR 8799, a 60-million-year-old star roughly twice as massive as our Sun, located 130 light-years away. The planets, captured with the help of adaptive optics on Gemini North, have estimated masses between seven and ten times that of Jupiter. “This is the first image of a multi-planet system, and these exoplanets are also the first at separations similar to Uranus and Neptune to be discovered by any means,” wrote Marois. Interestingly, the host star shows excess infrared emission—evidence of a dust belt located outside the new planets, somewhat akin to the Kuiper Belt beyond Neptune in our solar system. The researchers did not have spectra of the planets in hand, but were able to see their orbital motion around the star. That’s because the two outer planets were recovered in an image taken four years earlier with Keck, and the innermost (thus the fastest) one was seen to move over the course of a year. What’s more, David Lafrenière was able to recover the outermost planet in images taken ten years earlier with an infrared camera on Hubble. He downloaded the publicly available data from the Space Telescope Science Institute’s online archive, and applied advanced imaging processing algorithms to find the faint companion that had been missed by others. In late 2010, Marois and colleagues unveiled a fourth, innermost planet in the system. All four planets appear to orbit in the same plane, just like planets in our solar system, suggesting that they formed out of a disk. Some scientists, including Marois, think that the planets’ hefty masses and wide separations imply they formed rapidly, as a massive disk around the newborn star became gravitation-ally unstable and fragmented into pieces, which in turn acted as planet seeds.

The other planet image to be unveiled on the same day as the HR 8799 system came from a group led by Paul Kalas of the University of California at Berkeley. They reported an object no more than a few times more massive than Jupiter that appeared to shepherd the dust ring around Fomalhaut, also a 2-solar-mass star, with an age of about 200 million years just twenty-five light-years from the Sun. It was seen in images taken with Hubble at two optical wavelengths, and is located at about 100 times the Earth-Sun distance from Fomalhaut. The object is not detected yet at infrared wavelengths, and there is no spectrum in hand. It is possible that some or all of the light of Fomalhaut b in the optical images comes from a dust cloud, rather than a planet itself. More observations are clearly needed before one can be sure about the nature of this companion.

Figure 7.3. Image of the three planets orbiting the young star HR 8799. Credit: NRC-HIA, C. Marois, B. Macintosh, and Keck Observatory

The Fomalhaut discovery got the most media attention, perhaps because it was announced at a NASA press conference, though the “planet family portrait” of HR 8799 also received a fair amount of coverage. Some reporters, like Richard Harris of National Public Radio, put the findings in context. “Astronomers are getting their first real glimpses of planets in orbit around distant stars. Over the past decade, more than 300 other worldly worlds have been detected indirectly. . . . But the most recent planet discoveries are actual photo-ops. . . . There have been three reports in the past two months purporting to show images of planets in solar systems around nearby stars,” he explained. Asked for a comment, I told table 7.1 Directly Imaged Extrasolar Planets him: “Not only is it exciting just because we have pictures for the first time, but also because these pictures are revealing an entirely new population of planets that were not accessible to the previously used methods for planet detection.”

There was yet another announcement of a planet image before the year was out. In December 2008, a European team led by Anne-Marie Lagrange of the Grenoble Observatory in France reported a 9-Jupiter-mass planet candidate just 8 AU from the nearby star beta Pictoris. The star harbors a famous debris disk, seen close to edge-on, and over the years several researchers had suggested that planets embedded within it might account for the disk’s apparent twists. The researchers have now confirmed that the planet orbits the star, and are attempting to take spectra of it.

These first findings are mere harbingers of what is to come in the next decade. Two new instruments, specifically designed for planet imaging, will be mounted on the Gemini South telescope and the European Very Large Telescope, both in Chile, by 2012. Each has somewhat different niches and will survey several hundred nearby targets. Both instruments will take advantage of “extreme adaptive optics” to take sharper images than hitherto possible, and employ a host of other tricks to achieve the high contrast needed to detect dim planets next to bright stars. One trick is the use of coronagraphs. Invented by the French astronomer Bernard Lyot in 1930 to observe the Sun’s outer realms without having to wait for a solar eclipse, a coronagraph at its simplest is an occulting mask placed inside a telescope (or instrument) to block the bright, central part of the Sun. Modern designs use more sophisticated shapes for the mask, to improve the suppression of starlight while revealing extremely faint companions in the surrounding area. Coronagraphs will also be used for planet imaging with the James Webb Space Telescope, the 6.5-meter successor to Hubble, scheduled for launch in 2014. Even with these instruments, we will be limited to imaging giant planets, mostly around youngish stars in the solar neighborhood.

Faint Blue Dots

In principle, though, an advanced coronagraph could do much better. Installed in a precisely engineered, modest-size telescope in space, free from atmospheric blurring effects, it could take pictures of Earth-like planets around the nearest stars. That is no mean task. In visible light, an Earth twin would be nearly 10 billion times fainter than its star and separated from it by less than one-tenth of an arcsecond, an angle that is 20,000 times smaller than the apparent diameter of the Moon in the sky. John Trauger and Wesley Traub of the Jet Propulsion Laboratory in Pasadena, California, managed to achieve the necessary contrast with a lab apparatus in 2007, proving the technique’s potential. There are a few more engineering hurdles to overcome before the same can be done with a telescope in space. But Traub is hopeful: “If anything, it should be better, more stable in space,” he told me. Trauger and his collaborators as well as a team led by Olivier Guyon have already submitted mission concepts for consideration to NASA. Each team’s design calls for a telescope less than 2 meters in diameter—smaller than Hubble—but with high-precision optics and clever coronagraphs, all for a price tag of less than 800 million dollars.

Guyon, who splits his time between the Subaru Telescope in Hawaii and the University of Arizona, describes himself as “very much a hands-on person.” Growing up in the French countryside east of Paris, he discovered astronomy at the age of ten, thanks to a book given by a relative. His parents bought him a telescope a few years later. “Maybe they regretted it, because I spent many school nights outside,” he told me with a chuckle. “Being able to see with my own eyes the objects that I was reading about was really exciting for me. It made me want to become an astronomer,” he added. By age seventeen, he was building his first telescope. These days, he is putting together a lightweight telescope with a 1-meter mirror in his garage, which is equipped with an oven and a mirror-polishing apparatus. Once the telescope is completed, he plans to take it with him to the summit of Mauna Kea and to star parties with amateur astronomers. “Looking at the sky through the eyepiece of a telescope still gives me the most enjoyment. It’s almost a magical experience,” he said. He also enjoys working with frontline optics in his lab at the Subaru head quarters in Hilo.

As for imaging an Earth twin, “if we play it right, with an extremely stable setup, using the right approach, it can be done with a space telescope under two meters in size,” Guyon told me. “We have made amazing advances in optics and coronagraphs in the last decade,” he explained. “In the lab, things are working almost at the level we need. Perhaps it will take just a couple of more years to perfect the technology.” His mission concept, called PECO (for Pupil-mapping Exoplanet Coronographic Observer), calls for a 1.4-meter telescope. Trauger’s Eclipse mission design is a bit larger, at 1.8 meters. “Such a telescope will give us a good shot at imaging super-Earths, but an Earth twin would be somewhat of a long shot,” Guyon said. Either mission can capture the picture of an Earth-like planet only if it orbits one of the ten nearest stars. “The brightness of the planet will tell us something about its size. Changes in its color and brightness over time could tell us about clouds and weather, and reveal the length of its day. So we can learn a lot from direct images of a terrestrial planet, but you would need a bigger telescope to look for evidence of life,” he added. The proposed missions “will also reveal many giant planets and dusty disks, and prepare us for the next step—a larger space telescope optimized for planet imaging,” he pointed out. Wesley Traub of JPL agreed: “These small missions can image Jupiters and perhaps, if we are very lucky, see an Earth. They will certainly be useful for testing technology.” As he explained, “If you want to be sure of imaging several Earth-like planets, you need a larger space telescope to search the nearest 100 stars.”

There is another path toward imaging planets as small as the Earth. Interferometry, as it is called, permits astronomers to combine, or “interfere,” light waves collected by two or more telescopes to achieve far greater angular resolution than otherwise possible. Pictures taken by interferometers let us see fine details that single telescopes cannot and offer the prospect of separating a planet’s light from that of the bright parent star next to it.

The technique exploits the wave nature of light. If two waves identical in wavelength and amplitude overlap so that the crests of one wave coincide with the crests of the other, the waves amplify each other and produce a wave with twice the amplitude. But a small change will produce a very different result: If you shift one wave by half a wavelength with respect to the other, so that the crests of the first coincide with the troughs of the second, then the two waves cancel each other. Light plus light adds up to darkness. Because the wavelength of visible light is much shorter than the dimensions of most everyday objects—about 100 wavelengths fit across a dust speck—we rarely notice these subtle effects. When you see an array of shimmering rainbow colors on a puddle in the street, you’re seeing the result of interference, the cancellation of light waves on a thin film of oil on the water’s surface.

Even though optical interferometry is only now coming of age, its origins date back nearly a century to the efforts of Albert Michelson. Best known for measuring the speed of light, he won a Nobel Prize in Physics in 1907. Michelson knew that the angular resolution of a telescope—its ability to distinguish two stars that appear very close to each other—depended only on the size of the primary lens or mirror. If we double the size, we double the resolution. He realized that he could achieve the same effect by combining light from two smaller mirrors, without having to build a larger mirror. So, in the 1920s, he set out to “enlarge” the 100-inch Hooker telescope on Mount Wilson in southern California by adding an extension bar across its front end. This structure supported two small, adjustable mirrors separated by about 5 meters (200 inches). In his setup, it’s these mirrors that collected the starlight, not the Hooker telescope, which simply held the contraption in place. By moving the two small mirrors closer together and farther apart, Michelson was able to adjust the resolution of his “interferometer” until it matched the angular size of Betelgeuse, a bright red-giant star in the Orion constellation. It was the first time anyone was able to directly measure the diameter of a star other than the Sun.

Interferometry has come a long way since Michelson’s early experiments, especially in radio astronomy. Because the wavelengths of radio waves are typically 10,000 times longer than those of visible light, it’s a lot easier to work with them. The Very Large Array (VLA) consists of twenty-seven individual antennas laid out in a Y pattern in the New Mexico desert. For more than a quarter of a century, the VLA has allowed radio astronomers to see detail in objects that optical astronomers could only dream about.

Optical and infrared interferometers allow astronomers to achieve high angular resolution without requiring large, expensive mirrors. But detecting an extrasolar planet requires more than just superb resolution. Not only would the planet be very close to its parent star, but the much brighter star would overwhelm it. So we need to suppress the starlight enough to reveal the faint planet’s presence. Interferometers accomplish this with a clever trick: when combining light from two telescopes, if the light path from one is offset by half a wavelength relative to the other for the star’s location, its light would be canceled, or nulled. However, light coming from a slightly different direction, say from an orbiting planet, would not be canceled—letting us see the planet by minimizing the star’s glare.

While this challenging goal would require a future interferometer in space, ground-based efforts are already providing interesting science results while also contributing to technology development. The European Southern Observatory’s Very Large Telescope in Chile, consisting of four 8.2-meter units, has been used as an interferometer to measure sizes of nearby red dwarf stars, for example. Scientists are using the 10-meter Keck twins in Hawaii to test how well a star can be “nulled” to look for a dust disk and planets around it. The Large Binocular Telescope, recently completed in Arizona, was designed from the start for regular use as an interferometer.

A space version is still some years away. NASA’s Space Interferometry Mission (SIM) was to be the first long-baseline (10-meter) optical array in space, but its construction is currently on hold. Working high above the distorting effects of Earth’s atmosphere, SIM will be able to detect terrestrial planets indirectly through the astrometric motion they induce in their parent stars, although it will not have the sensitivity to detect light from extrasolar planets directly. Imaging of small planets will have to wait for the launch of NASA’s Terrestrial Planet Finder (TPF) or ESA’s Darwin, perhaps a decade from now. Darwin’s design calls for a flotilla of four or five free-flying 4-meter telescopes that will act as a nulling interferometer at mid-infrared wavelengths where the contrast between stars and planets is better. At the moment, TPF is also on hold, while Darwin is in development, though without a specific launch date.

So it will be a little while before we are able to take even a fuzzy picture of another Earth. Astronomers hope that discoveries of dozens of habitable terrestrial planets in the next few years will spur the political will and the financial resources to make it happen sooner rather than later.

¹ A revised distance measurement puts the system at 180 light-years and thus the pair’s separation at 40 AU.