Release 9 of 21

Astronomers Ponder Lack of Planets in Globular Cluster

Release date: Oct 31, 2000 1:00 AM (EST)
Astronomers Ponder Lack of Planets in Globular Cluster

Astronomers using the Hubble telescope made the first broad search for planets far beyond our local stellar neighborhood. They trained Hubble's "eagle eye" for eight days on a swarm of 35,000 stars in 47 Tucanae, located in the southern constellation Tucana. The researchers expected to find 17 "extrasolar" planets. To their surprise, they found none. These results may be the first evidence that conditions for planet formation and evolution are different in other regions of our Milky Way Galaxy.

The Full Story
Release date: Oct 31, 2000
Astronomers Ponder Lack of Planets in Globular Cluster

In the first attempt to systematically search for "extrasolar" planets far beyond our local stellar neighborhood, astronomers probed the heart of a distant globular star cluster and were surprised to come up with a score of "zero".

To the fascination and puzzlement of planet-searching astronomers, the results offer a sobering counterpoint to the flurry of planet discoveries announced over the previous months.

"This could be the first tantalizing evidence that conditions for planet formation and evolution may be fundamentally different elsewhere in the galaxy," says Mario Livio of the Space Telescope Science Institute (STScI) in Baltimore, MD.

The bold and innovative observation pushed NASA Hubble Space Telescope's capabilities to its limits, simultaneously scanning for small changes in the light from 35,000 stars in the globular star cluster 47 Tucanae, located 15,000 light-years (4 kiloparsecs) away in the southern constellation Tucana.

Hubble researchers caution that the finding must be tempered by the fact that some astronomers always considered the ancient globular cluster an unlikely abode for planets for a variety of reasons. Specifically, the cluster has a deficiency of heavier elements that may be needed for building planets. If this is the case, then planets may have formed later in the universe's evolution, when stars were richer in heavier elements. Correspondingly, life as we know it may have appeared later rather than sooner in the universe.

Another caveat is that Hubble searched for a specific type of planet called a "hot Jupiter," which is considered an oddball among some planet experts. The results do not rule out the possibility that 47 Tucanae could contain normal solar systems like ours, which Hubble could not have detected. But even if that's the case, the "null" result implies there is still something fundamentally different between the way planets are made in our own neighborhood and how they are made in the cluster.

Hubble couldn't directly view the planets, but instead employed a powerful search technique where the telescope measures the slight dimming of a star due to the passage of a planet in front of it, an event called a transit. The planet would have to be a bit larger than Jupiter to block enough light - about one percent - to be measurable by Hubble; Earth-like planets are too small.

However, an outside observer would have to watch our Sun for as long as 12 years before ever having a chance of seeing Jupiter briefly transit the Sun's face. The Hubble observation was capable of only catching those planetary transits that happen every few days. This would happen if the planet were in an orbit less than 1/20 Earth's distance from the Sun, placing it even closer to the star than the scorched planet Mercury - hence the name "hot Jupiter."

Why expect to find such a weird planet in the first place?

Based on radial-velocity surveys from ground-based telescopes, which measure the slight wobble in a star due to the small tug of an unseen companion, astronomers have found nine hot Jupiters in our local stellar neighborhood. Statistically this means one percent of all stars should have such planets. It's estimated that the orbits of 10 percent of these planets are tilted edge-on to Earth and so transit the face of their star.

In 1999, the first observation of a transiting planet was made by ground-based telescopes. The planet, with a 3.5-day period, had previously been detected by radial-velocity surveys, but this was a unique, independent confirmation. In a separate program to study a planet in these revealing circumstances, Ron Gilliland (STScI) and lead investigator Tim Brown (National Center for Atmospheric Research, Boulder, CO) demonstrated Hubble's exquisite ability to do precise photometry - the measurement of brightness and brightness changes in a star's light - by also looking at the planet. The Hubble data were so good they could look for evidence of rings or Earth-sized moons, if they existed.

But to discover new planets by transits, Gilliland had to crowd a lot of stars into Hubble's narrow field of view. The ideal target was the magnificent southern globular star cluster 47 Tucanae, one of the closest clusters to Earth. Within a single Hubble picture Gilliland could observe 35,000 stars at once. Like making a time-lapse movie, he had to take sequential snapshots of the cluster, looking for a telltale dimming of a star and recording any light curve that would be the true signature of a planet.

Based on statistics from a sampling of planets in our local stellar neighborhood, Gilliland and his co-investigators reasoned that 1 out of 1,000 stars in the globular cluster should have planets that transit once every few days. They predicted that Hubble should discover 17 hot Jupiter-class planets.

To catch a planet in a several-day orbit, Gilliland had Hubble's "eagle eye" trained on the cluster for eight consecutive days. The result was the most data-intensive observation ever done by Hubble. STScI archived over 1,300 exposures during the observation. Gilliland and Brown sifted through the results and came up with 100 variable stars, some of them eclipsing binaries where the companion is a star and not a planet. But none of them had the characteristic light curve that would be the signature of an extrasolar planet.

There are a variety of reasons the globular cluster environment may inhibit planet formation. 47 Tucanae is old and so is deficient in the heavier elements, which were formed later in the universe through the nucleosynthesis of heavier elements in the cores of first-generation stars. Planet surveys show that within 100 light-years of the Sun, heavy-element-rich stars are far more likely to harbor a hot Jupiter than heavy-element-poor stars. However, this is a chicken and egg puzzle because some theoreticians say that the heavy-element composition of a star may be enhanced after if it makes Jupiter-like planets and then swallows them as the planet orbit spirals into the star.

The stars are so tightly compacted in the core of the cluster - being separated by 1/100th the distance between our Sun and the next nearest star - that gravitational tidal effects may strip nascent planets from their parent stars. Also, the high stellar density could disturb the subsequent migration of the planet inward, which parks the hot Jupiters close to the star.

Another possibility is that a torrent of ultraviolet light from the earliest and biggest stars, which formed in the cluster billions of years ago may have boiled away fragile embryonic dust disks out of which planets would have formed.

These results will be published in The Astrophysical Journal Letters in December. Follow-up observations are needed to determine whether it is the initial conditions associated with planet birth or subsequent influences on evolution in this heavy-element-poor, crowded environment that led to an absence of planets.


In 1999, Ron Gilliland and colleagues tackled an ambitious eight-day planet-hunting marathon: Using NASA's Hubble Space Telescope, the astronomers studied a "city" of 35,000 tightly packed old stars for evidence of planets outside our solar system.

Gilliland, of the Space Telescope Science Institute in Baltimore, MD, went hunting for "extrasolar" planets in a nearby swarm of about a million stars, a globular cluster called 47 Tucanae in the constellation Tucana. He didn't use the Hubble telescope to snap pictures of planets. They're too small and dim to be imaged by any current observatory. Rather, he used Hubble's Wide Field and Planetary Camera 2 to watch for a subtle dip in the brightness of a star, an indication that a planet was passing in front of it. This event, called a transit, is similar to the moon eclipsing the Sun.

Gilliland came up empty-handed. But his sweeping survey shows dramatically how planet hunting has changed within the past five years. Until 1995, astronomers had no convincing proof that extrasolar planets around normal stars existed at all. But the advent of unique and efficient planet-hunting techniques helped catapult extrasolar planets from the realm of science fiction to reality. Astronomers embarked on this quest for extrasolar planets to solve some fundamental questions about our universe, such as, are there other life-sustaining worlds?

The planet hunt took off in late 1995 when a Swiss team, led by Michel Mayor, detected an unseen planet circuiting a Sun-like star. Geoffrey Marcy of the University of California, Berkeley, and Paul Butler of the Carnegie Institution of Washington quickly followed up that discovery with several of their own. Using new tools such as the Keck 10-meter telescope in Hawaii and its High Resolution Spectrograph, the community of planet hunters has now nabbed about 50 planets.

Astronomers have discovered these planets by watching for a telltale wobble in a star's motion. Called the radial-velocity method, it works like this: A planet exerts a gravitational tug on a star, which is detected from the star's slight change in speed as it moves toward and away from Earth. Astronomers measure this change by analyzing the star's light. This technique allows astronomers to determine the planet's minimum mass and the time it takes to complete one orbit around the star.

None of the planets found so far is a life-sustaining, terrestrial planet like Earth. The radial-velocity technique is biased toward finding large planets in short-period orbits. Most of them are bloated balls of gas, larger than Jupiter. A few, found recently are only as large as Saturn, the second largest planet in our solar system. Some of these celestial bodies are nearly as far from their host stars as Jupiter is from the Sun. Many others snuggle perilously close to their parent stars, completing an orbit in three or four days. They're nearer to their stars than Mercury is to the Sun. (Mercury is the closest planet to the Sun.)

Although astronomers kept ringing up planet after planet, skeptics wondered whether they were real. There were no pictures of these "ghostly" bodies, and their existence had not been confirmed by another observational method.

Then, in November 1999, astronomers hit the jackpot. Using two observational techniques, they confirmed the existence of a planet circling the star HD 209458, located 153 light-years from Earth in the constellation Pegasus. Astronomers used the radial-velocity method to discover the planet. From that observation, they calculated that the planet completed a circuit around the star every 3.5 days. They also figured out when the planet would transit the star. That's because the planet was so close to its parent star that there was a 10 percent chance that its orbit would be tilted edge-on to Earth. If so, ground-based telescopes and orbiting observatories would detect the planet crossing in front of the star.

Two teams of astronomers using ground-based telescopes recorded for the first time a planet moving across the face of the star. The transit, which lasted for about three hours, caused a 1.7 percent dip in brightness. Astronomer Geoffrey Marcy compared a transit to a bug flying in front of a light source, explaining that the bug blocks some of the lamp's light.

The transit technique furnishes astronomers with a planet's key diagnostic information, such as size, mass, and density. Since the size of the stars is known, astronomers can measure the size of the planets based on the percentage of blocked starlight. Then they can calculate the planet's mass and density. A planet's density tells astronomers whether it's rocky like Earth or a gas giant like Jupiter.

From calculations, the planet is 35 percent larger than Jupiter, 63 percent as massive as the solar system giant, and less dense than water. It is a hot, fluffy gas giant that is heated by its closeness to the star.

"We now have much more complete information on one planet," says Gilliland. "Until this transit observation, astronomers suspected it was a planet, but they had no physical proof, especially because of its short orbital period. Astronomers believed it was massive like Jupiter and was therefore a gas giant. But it could have been a large rock. This information will help astronomers understand how the planet formed and is evolving."

Although the transit method yields many of a planet's secrets, it does have its limitations. Only about 1 in 1,000 nearby stars with planets in short, three- to five-day orbits will transit between their stars and Earth. Astronomers, therefore, searching for planets using the transit method must observe thousands of stars to have any chance of success. That's why Gilliland and his team studied 35,000 stars in 47 Tucanae for planets. He employed Hubble's wide field camera with its superb resolution to analyze tens of thousands of stars at once.

The Hubble telescope also was used to study the confirmed planet around HD 209458. In April and May 2000, one of the teams that had confirmed the existence of the planet using the transit technique turned Hubble's "eagle eye" on the celestial body to look for the presence of rings like Saturn's and moons as large as our planet Earth.

Led by Tim Brown of the National Center for Atmospheric Research in Boulder, CO, the team used Hubble to watch the planet again transit the star because they needed even more precise measurements. In searching for moons and rings, the team was looking for tiny changes in starlight, which would be difficult for ground-based telescopes to distinguish. A planet passing in front of a star causes less than a 2 percent dip in brightness, but a much smaller moon blocks far less light. A moon the size of the Earth would block only about 0.01 percent of the starlight. A system of rings would lead to subtle changes in the slope of the light curve.

Although the team found no rings or moons, they showed that Hubble's location above Earth's atmosphere allows it to make precise measurements of even the slightest changes in a star's brightness. Based on the observation, the team produced a nearly flawless light curve of the transit.

"Hubble gives you much higher precision than you can get from the ground," Gilliland explains. "You can't look for the presence of rings or moons from the ground because starlight is disturbed too much due to atmospheric turbulence."