Debra Fischer of Yale University is a veteran planet hunter. In 1997, while still a graduate student at U.C. Santa Cruz, Fischer joined the Planet Search at the Lick Observatory in California and began scouring the skies for worlds orbiting faraway stars. At the time only two years had passed since the discovery of the first exoplanet, and the science of planet-hunting was in its infancy. The search required scientists to operate on the very edge of the sensitivity of their instruments, and even so only the easiest pickings could be detected at all. But Fischer was undeterred: sensing an opportunity to be part of a field that could transform humans’ understanding of their place in the universe, she jumped in with both feet. She became a key member of the search team, and soon took part in the discovery of several new exoplanets. From that time on, her list of discoveries just kept growing and growing.
Much has happened in the field of planet hunting since those early days. The first crop of planets found were oddballs – gas giants orbiting very close to their star, or “hot Jupiters” as they came to be known. But as the years went by, and techniques and instrumentation improved, smaller planets orbiting at greater distances could be detected. Furthermore the radial velocity technique – Fischer’s specialty – was joined by other methods, such as transit photometry and microlensing, each with its own unique advantages. As a result astronomers today are beginning to see not only distant planets, but entire planetary systems, and are gaining a deeper understanding of their dynamics and their prevalence. Fischer, now an acknowledged leader in the field, has moved east and now heads her own planet-hunting team at Yale.
Once again, however, the search for exoplanets is at a crossroads. As the sensitivity of detection instruments has increased, radial velocity searches have been able to detect planets as small as 4, and sometimes even 2 Earth masses. This is an enormous advance over the early days, but improving on it has proven to be extremely challenging. Despite astronomers’ best efforts, the holy grail of the exoplanet search, the detection of Earth-mass planets in the depths of space, remains just beyond reach. To Fischer and her team, this was a challenge they could not refuse.
The Quest for Precision
In their quest to detect a true exo-Earth Fischer and her colleagues settled on a simple strategy: to improve the precision of the technique they specialize in—radial velocity searches. In this type of search scientists use large telescopes to monitor the light spectrum of stars suspected of harboring planets. If a star’s spectrum shifts ever so slightly to the blue, and then towards the red, at fixed and regular intervals, it is almost certain that the star is moving—rocking back and forth along the line of sight from Earth. This movement is a strong indication that the star is orbited by another body, possibly a planet, whose gravitational tug causes the star to rock back and forth. Naturally, a massive planet is easier to detect than a low-mass planet, because it causes the star to move more rapidly and therefore creates sharper shifts in its spectrum. But over time planet-hunters have improved their measurements to the point where they can now detect a star’s motion even if it is as slow as 1 meter per second—the pace of a leisurely stroll. The planets that cause this rate of movement have a mass of only 2 to 4 Earths, making them among the lowest-mass planets discovered to date. But in order to find true Earth-mass planets astronomers will have to detect the motions of stars moving at less than 1 meter per second, and that has proven to be extremely difficult.
According to Fischer, there are two ways of breaking through the 1 m/s barrier. One option is to build specialized equipment dedicated exclusively to the search for exoplanets. European planet-hunters have, in fact, followed this route in building the dedicated HARPS spectrograph, housed at the European Southern Observatory’s facility in La Silla, Chile. Built especially to search for faraway worlds, it is the most sensitive spectrograph currently in use by planet hunters. But valuable though they are, instruments like HARPS are prohibitively expensive. Building a new and improved version to gain another meter per second edge over existing measurements will cost millions of dollars, and will still take years to complete.
So Fischer and her colleagues opted for a different approach: get creative, and try to squeeze every last drop of precision out of existing instruments. They came up with a new concept that they called "Fiber-optic Improved Next generation Doppler Search for Exo-Earths," or simply “FINDS Exo-Earths.” Then they called The Planetary Society to ask for funding. We knew a good investment when we saw it: we said yes.
A Simple Solution to a Difficult Problem
To understand how FINDS Exo-Earths works, consider the basic set-up of the radial velocity exoplanet search. First comes the telescope, collecting the light from the observed star. The light is then passed through a narrow slit, which lets only light from the star itself, and not from other objects, pass through. Finally the light from the slit is passed on to a highly sensitive spectrometer, where any spectrum fluctuations are recorded. The lens at the entrance to the spectrometer, where the light is concentrated, is called the "pupil." This simple set-up has worked extremely well in the past, and has led to the detection of hundreds of exoplanets. But to Fischer and her colleagues this was not enough: the light that passes through the slit, they found, illuminates the spectrograph pupil in slightly different ways at different times. This variation introduces a certain error into the spectrum measurements, and although it is miniscule, at the level of precision sought by Fischer’s team it is nonetheless significant. In fact, apart from the precision of the instruments themselves, it is the largest source of error left in radial velocity measurements. FINDS Exo-Earths works by correcting precisely this error.
As Fischer tells it, there are various reasons for the changing illumination of the spectrometer. Changes in the atmospheric temperature or pressure through the night can cause drifts in the telescope focus. Of more concern are changes over a span of seconds or minutes, which occur constantly. These might be caused by gusts of wind which shake the telescope or perturb the atmosphere, but the one common culprit is the guiding system of the telescope itself. As the star makes its way across the night sky, the guiding system is programmed to keep it always in the telescope’s sights, shining directly onto the spectrometer through the center of the slit. Even the slightest guiding error might cause the star to be positioned momentarily not at the center, but at the right or left edge of the slit. While the starlight is still collected when the star is off-center, the illumination pattern shifts, introducing an error into the spectrum measurements.
How does one correct for such an unpredictable error? Gusts of wind will always be there, and guiding mechanisms will always have some degree of error, however small. So Fischer and her colleagues correct for this by placing an optical fiber between the slit and the spectrograph. In this set-up the light from the telescope that passes through the slit would first pass through the optical fiber before continuing on to the spectrograph’s pupil. The fiber, they reasoned, would scramble the light from the slit, so that the precise position of the star would no longer matter. In place of the unsteady light beam that reacted to the slightest shift in the position of the star, the optical fiber would shine a smooth and even cone of light onto the spectrograph. This would eliminate the most troublesome source of error in radial velocity searches and increase measurement precision. By how much exactly remained to be seen, but Fischer and her colleagues were hopeful that in the long run it would make it possible to detect the movement of stars rocking back and forth at less than 1 meter per second.
Testing an Idea
The job of designing and building the optical fiber instrument fell to Julian Spronck, a postdoctoral scholar on Fischer’s team who specializes in optics. In addition to the 100 micron diameter optical fiber itself, Spronck’s prototype included an elaborate set of mirrors and lenses. This, explained Spronck, is to ensure that the light enters and exits the fiber at precisely the right angles, since even the slightest deviation from the specified angles would cause much of the starlight collected by the telescope to be lost. When completed, the instrument measured 2 x 1 feet and fitted smoothly onto the Hamilton spectrograph—the one used for radial velocity measurements at the Lick Observatory in California. The FINDS prototype has been in place at the Lick Observatory for several months now, and the results are impressive. The illumination of the spectrograph, previously so sensitive to every gust of wind or tiny tracking irregularities, now seems nearly immune to such errors. Fischer and her team know this, because they have been tracking the "point spread function" (PSF) of the pupil, which indicates what a point of light, such as a star, looks like after it gets spread out by an instrument’s optics. Strong variation in the PSF indicates that the illumination is unsteady, as it was before Spronck’s visit to the Lick. But with FINDS in place the PSF is more that 10 times steadier than it was in the past, indicating that the light that passes through the optical fiber shines on the pupil in a steady even cone, regardless of the precise position of the star. And that is precisely what FINDS was designed to accomplish.
With the unpredictable fluctuations in illumination removed from the spectrograph’s readings, the next step was to see by how much this improved actual radial velocity readings. To test this, the team pointed the Lick telescope at 3 stars whose radial velocity motion is known with a high degree of accuracy, and measured their motion repeatedly. Inevitably, any single reading of a star’s motion has a certain degree of error, measured by how far the actual reading is from the star’s true velocity. By comparing dozens of measurements of the 3 stars with and without the FINDS, Fischer and Spronck showed that the instrument reduces error by an average of 30%. "This is dramatic,” Fischer said. “We usually fight to squeeze out improvements of a few percent."
Spronck is currently working to improve the FINDS prototype, trying out different types of optical fiber and different diameters. He is now starting to experiment with fibers that have exotic shapes: squares, hexagons and octagons, rather than circular cross-sections, because these special fibers have better scrambling properties. For the time being Spronck has to be personally present at the Lick Observatory whenever the instrument is used, since he is the only one who can calibrate the lenses and mirrors with sufficient precision. Once the permanent instrument is built and installed, it will require no special attention and will be maintained by the observatory’s technical staff. Furthermore, the spectrograph measurements will then be calibrated to the precise type of fiber used, resulting in even better precision than was achieved with the prototype.
On to the Big Leagues
While perfecting the prototype at the Lick Observatory, Fischer and her team are also intent on trying their instrument at the Keck Observatory in Mauna Kea, Hawaii. The Keck’s twin 10 meter reflectors are among the most powerful telescopes in the world, and they have been used to discover some of the lowest mass planets yet detected. If FINDS is to fulfill its promise of detecting Earth-like planets, it will do so at giants like the Keck, and it was important to try out the instrument there. As it turned out, there was no room at the Keck for the instrument Spronck had built for the Lick Observatory, so Spronck had to go back to the drawing board. Ultimately, he designed and built a miniature "matchbox FINDS," measuring only 1 x 3 inches where the Lick instrument measured 2 x 1 feet. On the night of September 29, 2010, Fischer and Spronck had their chance to try out the matchbox design at the Keck. They compared 25 measurements taken with FINDS that night with 147 measurements taken without it over the previous 6 years without the instrument. The results were striking: FINDS improved the PSF stability—i.e. the stability of the spectrograph illumination—by a factor of 10!
Like all good things, the improvement in precision achieved by FINDS comes at a price. Inevitably, some of the light gathered by the telescope is lost when it is passed through an optical fiber instead of being sent directly to the spectrograph. This loss is hard on astronomers, who are trained to gather every possible photon collected by their telescopes. Nevertheless, according to Fischer, a certain loss is acceptable, and is more than compensated for by the increased precision. Currently the light losses at the Lick are around 20%, which is close to the goal Fischer has set.
At the Keck, however, where FINDS could ultimately be used to search for exo-Earths, the results so far have not been what Fischer and Spronck had hoped for. Spronck’s prototype "matchbox" FINDS fit snugly within the space available at the Keck and effectively stabilizes the spectrograph illumination, but it also lost almost 60% of the light gathered by the telescope. And that, says Fischer, is simply too high a price to pay.
Fischer and Spronck are currently hard at work on a new design for the Keck, which will preserve the light gathered by the telescope and still stabilize the illumination of the spectrograph. To do this, however, they need more funding, and once again they asked The Planetary Society to help.
FINDS Exo-Earths has already shown that it is a difference maker in the race to discover faraway Earths. As prototypes are replaced by permanent instruments, its already impressive results will continue to improve, leading to the detection of lower and lower mass planets. It is, says Fischer, an important technological step towards the next generation of planet-hunting spectrographs. But FINDS is also something else: a creative low-cost solution to a persistent problem that has challenged even the best funded groups using the most expensive and sophisticated instruments. It is a project in which a modest investment can reap huge scientific and technological dividends. It is, in other words, what The Planetary Society does best.