Two teams of scientists keep a constant watch on the southern sky for such events. OGLE (Optical Gravitational Lensing Experiment) is a collaboration of Polish and American astronomers observing in Chile; the other, MOA (Microlensing Observations in Astrophysics), is a collaboration between astronomers in New Zealand and Japan. Together OGLE and MOA monitor some 10 million stars, night after night. As soon as they spot a microlensing event, indicated by a star that appears suddenly to brighten, they flash an alert to PLANET (the Probing Lensing Anomalies NETwork). The astronomers in the PLANET collaboration, who are spread out among four telescopes throughout the southern hemisphere, then begin monitoring the events intensively. A typical microlensing event lasts for many weeks, but the part that could reveal a planet lasts only a few hour to a few days - and there's no way to predict when the interesting part will occur.
Bumps and Dips
All stars are constantly moving. And, occasionally, as viewed from Earth, one relatively close star moves in front of another, far-more-distant star. When this occurs, the gravitational tug of the closer star bends and focuses the light of the more-distant star, just like a magnifying glass. Not enough to let us see any significant detail, but enough to make a normally very dim star appear measurably brighter. It is these changes in brightness that Sackett and her colleagues study.
It's not the distant star that astronomers care about; it's just serving as a big light bulb. The closer star, the one that acts as the lens, is the interesting one, the one that researchers hope will reveal a planetary companion.
When the lensing star has no planet (or when its planets are too small, or too close, or too far away, or not lined up just so), the change in the distant star's brightness will be seen to slowly rise, and then slowly fall again. A graph of this change in brightness over time, known as a light curve, has a simple bell shape.
If, however, the lensing star has a planet, the planet produces a "defect" in the gravitational lens, like a bubble does in a glass lens. This region of the lens focuses light slightly differently than the rest of the lens.
If the planet is the right size (at least as big as Saturn) and the right distance from the lensing star (between 1 and 5 AU - 1 AU is the distance between the Sun and the Earth), and if the distant star passes close enough to the gravitational "bubble" caused by the planet, the resulting light curve will have an unusual bump or dip in it. >From the height and width of the dip, researchers can learn something about the size and orbit of the planet.
Unfortunately, although 43 microlensing events have been fully characterized to date, no planets have yet been detected. That doesn't mean that nothing has been learned, however. The absence of bumps and dips also provides useful information. Statistically speaking, Sackett says, she and her colleagues are certain that no more than half of the 43 lensing stars studied host a Jupiter-sized planet with a Jupiter-like (5 AU) orbit. One can't say for certain that none of the lensing stars have a Jupiter-like planet, Sackett explains. It's possible that some do, but that the distant stars didn't pass behind the planet-induced defects in the lensing stars.
"This," says Sackett, "is very much about being an astronomer, where you never get to design an experiment. If I could design an experiment, I would arrange to have the background star pass exactly where I like. You can't be a control freak in astronomy."
There are other shortcomings of the microlensing technique. One is that a microlensing event, by its random nature, is a one-shot affair. It happens; you watch it; it's over. There is no second chance to verify the results, because the particular alignment of the distant star, the lensing star and Earth, which created the event, is never repeated. But because PLANET involves astronomers in different locations observing the same event, they can correlate their results.
That's not quite as bad as it sounds. The majority of stars in the galaxy (about 80%) fall within a narrow range of masses, with M dwarfs on one end of the range and Sun-type stars, about 3 times as massive as the M dwarfs, on the other. So it follows that most microlensing events are caused by these abundant types of stars. It's just not possible - yet - to say for certain what type of star is responsible for any particular event.
As Sackett explains, "Microlensing is like studying populations of people, rather than like studying individuals. We don't actually know, on an event-by-event basis, what kind of star is producing the lensing event. We know a lot about the kinds of stars between here and the galactic center, where we're looking. So statistically, we can tell you there must be this percentage that are M dwarfs and that percentage that are solar-type stars, and so forth. But if you say, 'Okay, which ones are the solar-type stars; show them to me,' I can't."
Nevertheless, as the microlensing technique begins to catalog planets, or fails to detect them, it will be possible to build up a statistical picture that will help astronomers understand planetary systems around a range of different types of stars. In particular, because M dwarfs are the most common stars in the galaxy and therefore are responsible for the majority of lensing events, microlensing searches will provide insight into M-dwarf systems. Planet searches using the radial-velocity technique are just beginning to look at M dwarfs.
Scientists who use the two different techniques are looking forward to the day - although it may be some years away - when their research overlaps and they can compare notes. Their overarching goal is to learn how planets form around different types of stars, and how they migrate inward - if, indeed, they do - after they form. This, in turn, should help to understand the likelihood of finding inhabited planets in the galaxy.
According to Sackett, it may be possible to identify the precise mass of a lensing star (and its planet). Stars of different masses radiate with distinct spectra. By identifying the spectrum of a lensing star, then, one could, in theory, determine its mass. Moreover, for a star with a planet, because the relative mass of the star and planet would be known, the planet's mass could also be calculated.
The problem is that in a lensing event the distant star - the light bulb - far outshines the lensing star, whose light gets lost in the glare. Only the world's most powerful telescopes, such as the Keck scopes in Hawaii and the Very Large Telescopes (VLT) in Chile, are capable of extracting the dimmer star's spectra. But, says Sackett, "to be perfectly honest, I don't think they'll give us time on the Keck or VLT unless we have a planetary signal. They won't do it for just a random microlensing event."
Meanwhile, OGLE has expanded its work to include yet another type of planet-detection technique: transits. A transit occurs when (as viewed from Earth) a planet crosses the face of a star, causing its light to dim slightly. Because OGLE is already looking for changes in star brightness as an indicator of microlensing event, adding a search for transits is not that difficult to integrate. The transit technique is likely to be the first planet-hunting technique to reliably detect Earth-size planets. The first such detection could come within the next few years, when space observatories are launched to search for planetary transits.