Summary: When they turned the Hubble Space Telescope on a distant globular cluster of stars, astronomers expected to find fifteen or twenty planets. They found zero.
When they turned the Hubble Space Telescope on a distant globular cluster of stars, astronomers expected to find fifteen or twenty planets. They found zero.
At an age of about 10 billion years, globular clusters contain some of the oldest stars in the Milky Way. Scientists study these ancient, densely packed star systems to learn about the lives of stars, the formation of galaxies, and the evolution of the universe. By searching for planets in globular clusters, it is also possible to discover whether planets formed when our galaxy was relatively young.
Astronomer Ronald Gilliland of the Space Telescope Science Institute in Baltimore, Maryland, and his team searched for planets in the globular cluster 47 Tucanae, a crowded group of stars located about 15,000 light-years from Earth. They used the Hubble Space Telescope (HST) to monitor the brightness of more than 34,000 stars over a period of 8 days.
Gilliland and his team searched for "hot Jupiters" - planets approximately the size of Jupiter that orbit extremely close to their stars. While some experts consider hot Jupiters to be rare, one of every 100 stars searched locally using radial-velocity techniques has been found to have such a planet. Gilliland's colleagues expected to find 15 or 20 of these close-orbiting Jupiter-like planets in 47 Tucanae. Instead, they found none.
"Before the observations, I didn't really believe one way or the other whether planets exist in globular clusters such as 47 Tuc," says Gilliland. "I believe that if the type of planets - gas giants with 3- to 5-day orbits - exist with the frequency in 47 Tuc as found in local stars, then our HST experiment wo uld have seen these. So it seems that these short-period, gas-giant planets really are more rare in 47 Tuc."
The lack of "hot Jupiters" in 47 Tuc may point to a fundamental difference, when it comes to planet formation, between our own galactic neighborhood and globular clusters in distant parts of our galaxy. The results of Gilliland's study, however, do not rule out the possibility that 47 Tuc could contain other sorts of planets. Limits in the Hubble Space Telescope's technology - it was never designed to do this type of work - prevent it from detecting extrasolar planets much smaller than Jupiter.
"Given the paucity of observational knowledge about where we might find planets, the investment of a week of HST time to check out 47 Tuc made perfect sense, even if it seemed like a long shot at the time," says Alan Boss of the Carnegie Institution of Washington. "The fact that nothing was found is not surprising, given the many reasons why detectable hot Jupiters might not exist in a globular cluster. But at least now we know the answer."
How the Search was Carried Out
If a sufficiently large planet around a distant star crossed between the star and the HST, it would cause the star to dim perceptibly. The effect is similar to how an eclipse of our Sun would appear to us if our moon were many times smaller. The search for these signature dips in brightness is called the transit method of planet detection. While there are many other ways to detect extrasolar planets, this method was considered the best to use for 47 Tuc because of the great number of stars to be observed. Scientists can simply watch the stars over time and see which ones fade periodically.
The transit method has some limitations, however. The planets' orbits must be seen edge-on from Earth. If we view on orbiting system from either above or below the middle horizon, the planet will not block our view of the star, and the star's brightness will not appear to dip. Globular clusters have so many stars, however, that the odds of finding at least some edge-on orbits in those systems are considered to be quite good.
Another limitation of the transit method is time. A planet must be observed to pass in front of its star more than once in order for scientists to confirm the planet's existence. The time available to use the Hubble Space Telescope, however, is short - at 8 days, Gilliland's HST study was considered exceptionally long. Since very prolonged observations are currently difficult to perform, such studies will only find planets extremely close to their stars, with orbits of only a few days.
For all we know, most of the stars in 47 Tuc have planets; they simply weren't detectable with the method used. But what is the likelihood that globular clusters like 47 Tuc do contain planets?
"So far we only know that 47 Tuc doesn't seem to have any hot Jupiters," says Boss. "That is a long way from knowing that it does not have any other planets that might be interesting from the point of view of habitability. But a globular cluster is not the first place one would look to try to find a habitable planet."
Many scientists believe planets would not easily form in the globular cluster environment. While our Sun and nearby stars are all rich in planet-forming heavy elements like carbon, oxygen, and iron, globular clusters contain far lower concentrations of those elements.
"The locally detected planets seem to occur primarily around host stars with higher than average metal abundances," says Gilliland. (When astronomers speak of "metals" they include elements like carbon and oxygen.) "Forty-seven Tuc, although metal-rich by most globular-cluster standards, is metal-poor compared to the local sample of stars that have planets. Perhaps a higher metal fraction is required to have planets form."
Another barrier to planet formation is that the stars in globular clusters are densely packed. The immense gravitational tug of war among so many stars could disrupt or even destroy planets that try to form.
"We don't really know for sure why hot Jupiters don't seem to exist in globular clusters," says Boss. "The crowded environment is a likely contributor to the shortfall - it can result in stripping of the disks that form planets, either by irradiation or tidal effects during close encounters, and tidal effects could also scramble and eject any planets that did manage to form.
"It is conceivable that planets exist around the stars in globular clusters, though it will be some time before we can do much to detect them in such distant environments," Boss continues. "I don't know of any plans to build specialized instruments or satellites to try to detect such planets in the future. The focus is on finding the nearby ones first, which is more than difficult enough."
Gilliland says he has no further plans to search for planets in 47 Tucanae, or any other globular cluster, although he does not rule out the possibility entirely.
"Forty-seven Tuc was selected as the ideal globular cluster target, and I don't know of any others that would be as promising," says Gilliland. "If I knew of a way to effectively pursue planets further in 47 Tuc I might try, but do not see any means of doing so. I reserve the right, however, to keep thinking along these lines."
In the meantime, Gilliland is working to get NASA approval on the Kepler mission - a space telescope that would use the transit method to search for Earth-like planets around stars more similar to our own Sun than those in 47 Tuc.
"Unlike the 47 Tuc case, where the stars examined were not like those in our neighborhood, Kepler would observe relatively normal stars and firmly establish the frequency of Earth-like planets," says Gilliland. "The Kepler Mission holds the best near-term hope of studying extrasolar planets likely to be of interest for astrobiology endeavors."
Alan Boss agrees, but also thinks the combination of ground-based observations with projects like the Full-sky Astrometric Mapping Explorer (FAME), the Space Interferometry Mission (SIM), and the Terrestrial Planet Finder (TPF) will contribute to the discovery of extrasolar worlds.
"The next advances will continue to come from the ground-based programs, which are poised to discover Jupiter-like companions to nearby stars in the next decade, planets that may be signposts pointing to systems that may be analogs of our own solar system," says Boss. "Toward the end of and beyond this decade, the space-borne telescopes will begin to detect extrasolar planets: FAME; Kepler, if chosen; SIM; and TPFâthe latter three may be able to find the first extrasolar Earths. Then things will really get exciting!"