News Release

NCAR scientists find sodium in exoplanet atmosphere

Peer-Reviewed Publication

National Center for Atmospheric Research/University Corporation for Atmospheric Research

BOULDER--A recent student visitor to the National Center for Atmospheric Research and an NCAR senior scientist have made the first direct detection of the atmosphere of a planet orbiting a star outside our solar system and have obtained the first information about its chemical composition. Their unique observations from NASA's space-based Hubble telescope demonstrate that it is possible to measure the chemical makeup of extrasolar planetary atmospheres and potentially to search for chemical markers of life beyond Earth.

Lead investigator David Charbonneau, a former graduate student fellow at NCAR, and NCAR's Timothy Brown, used the Hubble telescope's spectrometer to detect the presence of sodium in the extrasolar planet's upper atmosphere. Charbonneau is now at the California Institute of Technology. The National Science Foundation, NCAR's primary sponsor, and NASA supported the research.

"Only a decade ago, planets outside the solar system were still in the realm of science fiction," says Charbonneau. "Searching for a star's unseen planetary companion was crazy. Hoping to see its atmosphere was even crazier." Now planets are discovered monthly, and even their atmospheres are in reach, he says. "Suddenly, discussing searches for Earth-like planets seems quite reasonable."

The graduate student and the senior scientist worked together at NCAR's High Altitude Observatory to develop their computer model and refine their observational approach. The model predicted which chemicals and how much of each to look for. To take advantage of the particular strengths of the Hubble spectrograph, they narrowed their search to sodium. Last month, two years after their initial proposal, they were rewarded when the anticipated thick black lines showed up at the predicted position in the yellow-green region of the starlight's visible spectrum--a sure sign of sodium atoms absorbing and scattering starlight. The Hubble's highly precise measurements confirmed that the sodium exists not only in the star's atmosphere but in the planet's as well.

"It's hugely exciting to nail down anything at all about something as mysterious as planets outside our solar system," says Brown. "Is this planet's chemical personality unique or is it typical of a certain class of extrasolar planets? We have no clue. We hope to find out." Extrasolar planetary atmospheres are tough to crack scientifically because the observable quantities are tiny and the measurements must be extremely precise, he says.

First color in a planet's portrait

The planet orbits a yellow, sun-like star called HD 209458, a seventh-magnitude star (visible with an amateur telescope), which lies 150 light-years away in the autumn constellation Pegasus. Its atmospheric composition was probed when the planet passed in front of its parent star, allowing the scientists for the first time ever to see light from the star filtered through the planet's atmosphere. Before their observation, there was no direct proof of the existence of an atmosphere on any of the 76 extrasolar planets discovered to date, let alone a measurement of specific chemical composition.

Charbonneau and Brown are currently using the Hubble's spectrograph to scrutinize faint starlight reflected off the planet toward Earth just as the planet passes behind the edge of its star. The results, based on a barely measurable .01% whisper of starlight, should reveal the color and reflectivity of the planet and indicate whether clouds of dust or metallic elements might be present.

"In some computer models, this planet is blacker than coal," says Brown. "In others, it's bright white, like Venus. Only observations can tell us what's real." As soon as next spring, artists may be using the scientists' results to paint the planet's portrait in true- to-life color.

From low-tech breakthrough to the Hubble

Charbonneau already earned a place in astronomy textbooks during his visit to NCAR. While working on his Ph.D. at the Harvard-Smithsonian Center for Astrophysics, he was selected as a Gordon Newkirk Graduate Student Fellow in NCAR's High Altitude Observatory from August 1999 to August 2000. In the fall of that year, he and Brown experimented with a small telescope, designed and fabricated at NCAR using Brown's homemade optics. Brown built it to test what was then an untried technique, called the transit method, to confirm the existence of the HD 209458 planet by observing the dimming of starlight as the planet passed between its star and the Earth.

The same planet had been discovered a few months earlier using a more common detection procedure based on its slight gravitational tug on the star. From that observation, the planet was estimated to be 70 percent the mass of the giant planet Jupiter (or 220 times more massive than Earth).

The low-tech telescope first searched for transits from a chicken coop outside Boulder, then from a makeshift shed next to an NCAR parking lot. (It now scans the skies for planet-bearing stars from its perch on Tenerife in the Canary Islands.)

By November 1999, Charbonneau and Brown had results. To date the planet circling HD 209458 remains the only extrasolar planet whose physical existence has been confirmed using the transit method of detection. Within a month of that success, the two scientists proposed using the Hubble spectrograph to identify elements in an extrasolar planetary atmosphere.

In the current Hubble experiment, the astronomers actually saw less sodium than predicted for the Jupiter-class planet. One interpretation suggests that high-altitude clouds in the alien atmosphere may have blocked some of the light.

The Hubble observation was not tuned to look for gases expected in a life-sustaining atmosphere (which is improbable for a planet as hot as the one observed). Nevertheless, this unique observing technique opens a new phase in the exploration of extrasolar planets, say astronomers. Such observations could potentially provide the first direct evidence for life beyond Earth by measuring unusual abundances of atmospheric gases caused by the presence of living organisms.

A gaseous giant

The planet is an ideal target for repeat observations because it transits its star every 3.5 days--the extremely short time it takes the planet to whirl around the star at a distance of merely 4 million miles from the star's searing surface. This precariously close proximity to the star heats the planet's atmosphere to a torrid 2,000 degrees Fahrenheit (1,100 degrees Celsius).

The team--which also included Robert Noyes of the Harvard-Smithsonian Center for Astrophysics and Ronald Gilliland of the Space Telescope Science Institute in Baltimore, Maryland--had previously used transit observations by Hubble and ground-based telescopes to confirm that the planet is primarily gaseous, rather than liquid or solid, because it has a density less than that of water. (Earth, a rocky rather than a gaseous planet, has an average density five times that of water.) These earlier observations thus established that the planet is a gas giant, like Jupiter and Saturn.

The planet's swift orbit allowed for observations of four separate transits to be made by Hubble in search of direct evidence of an atmosphere. Though the star also has sodium in its outer layers, the spectrograph precisely measured a very slight additional filtration of sodium (an enhancement of less than one percent) as the starlight passed through the planet's atmosphere.

The team next plans to look at HD 209458 again with Hubble in other colors of the star's spectrum to see which are filtered by the planet's atmosphere. They hope eventually to detect methane, water vapor, potassium, and other chemicals in the planet's atmosphere.

Hot Jupiters and the search for Earth-like planets

Once other transiting giants are found in the next few years, the team expects to characterize chemical differences among the atmospheres of these planets. These anticipated findings would ultimately help astronomers better understand a bizarre class of extrasolar planets discovered in recent years that are dubbed "hot Jupiters." They are the size of Jupiter but orbit closer to their stars than the tiny innermost planet Mercury in our solar system. While Mercury is a scorched, airless rock, these planets have enough gravity to hold onto their atmospheres, though some are hot enough to melt copper.

Conventional theory is that these giant planets could not have been born so close to their stars. Gravitational interactions with other planetary bodies or gravitational forces in a circumstellar disk must have carried these giants via spiraling orbits precariously close to their stars from their birthplace in cooler regions farther out, where they bulked up on gas and dust as they formed.

Proposed moderate-sized U.S. and European space telescopes could allow for the detection of many much smaller Earth-like planets by transit techniques within the next decade. The chances for detection will be more challenging, since detecting a planet orbiting at an Earth-like distance will mean a much tighter orbital alignment is needed for a transit. And the transits would be much less frequent for planets with an orbital period of a year, rather than days. Eventually, study of the atmosphere of these Earth-like planets will require meticulous measurements by future larger space telescopes.


Note to editors: David Charbonneau and Timothy Brown will participate in a national press conference to be held at 1:00 p.m. EST on Tuesday, November 27, 2001, at NASA Headquarters, 300 E Street, Washington, D.C. For more information, please contact Ray Villard, Space Telescope Science Institute, Baltimore, Maryland (410-338-4514,

On the Web: Visuals will be available November 27 at

Ray Villard
Space Telescope Science

Cheryl Dybas

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.