Another "First": Emission Spectra of Extrasolar Planets
22 Mar, 2007 11:21 am
The rapidly-evolving field of research in extrasolar planets (or exoplanets) has achieved another milestone, with the announcement of the first measured emission spectrum from two different planets orbiting other stars. The results revealed a surprise: the water that was expected to be in the atmospheres of these planets was nowhere to be found.
The measurements have been achieved using NASA's orbiting Spitzer Space Telescope, which carries three instruments designed to work in the infrared region of the spectrum. These observations made use of the Infrared Spectrograph (IRS), which provides wavelength coverage in the mid-infrared region of the spectrum. Spitzer was never designed for studying exoplanets; it is a tribute to the teams that designed and built Spitzer that it has the stability to perform such precise measurements.
The observations were also made possible by the specific geometry of these exoplanetary systems. Both are "transiting" planets, meaning that they cross directly in front of the stars that they orbit, as seen from Earth. So far, 14 exoplanets are known to exhibit transits. The transiting planets have provided the most detailed information to date on the nature of exoplanets. At the time of transit (planet in front of star), the measured stellar light decreases in proportion to the ratio of planetary to stellar area, and so a measurement of the transit curve provides a direct measurement of the radius, true mass, and thus the density of the exoplanet. This was done for the first transiting planet to be discovered, HD 209458 b (see Charbonneau et al 2000, Henry et al 2000).
If the planet crosses in front of the star, it must pass behind the star, roughly half an orbital period later (exactly half a period for a perfectly circular orbit). This disappearance of the planet behind the star is called the secondary eclipse, or the eclipse for short. In the visible region of the spectrum, it is not possible to detect the secondary eclipse, because the starlight reflected from the top of the planet's atmosphere is too small compared to the star itself. However, if we look in the infrared, because the planet is so hot, it has a significant energy output in its own right, which is a few tenths of a percent of the stellar output. A small signal, but detectable with a stable instrument.
Thus, the idea behind the observation is elegantly simple (although in practice its quite difficult to tease this information out of the data). The system (star plus planet) is observed outside of eclipse, just before the planet is set to disappear behind the star. (Note that the planet and star cannot be spatially separated--meaning it's not yet possible to take an image--because they are too close together.) Then the system is observed when the planet disappears behind the star. By subtracting the in-eclipse observations (star only) from the out-of-eclipse observations (star plus planet), we can in principle derive the energy output of the planet alone. This was first done at a single wavelength for two different planets (see Deming et al 2005, Charbonneau et al 2005), leading to the first observational measurement of the temperatures of these objects.
The new research results have taken the observation to the next level and conducted the observation at a range of wavelengths in the infrared, thereby determining an emission spectrum of the planet. A team led by Jeremy Richardson at NASA Goddard Space Flight Center studied the exoplanet HD 209458 b (see Richardson et al, Nature, 2007). They found no evidence for the water absorption feature (at roughly 8 - 9.5 microns) that was expected to appear. They further find a weak but real feature near 9.6 microns, which corresponds to the Si-O fundamental stretching mode. This is suggestive of silicate clouds. Such clouds, high in the atmosphere, could be thick enough to mask the water in the atmosphere below, and thus explain the lack of water detected in the spectrum.
A second team led by Carl Grillmair at Caltech found similar results for HD 189733 b (see Grillmair et al, ApJL 2007). Although they find no evidence for silicate emission, their analysis suggests the presence of strong winds in the planet's atmosphere, which would be one mechanism by which the strong radiation from the star hitting the day side of the planet could be transmitted to the night side of the planet.
It is worth pointing out that there are other explanations for the lack of water in the observed spectra. These include the situation noted above, namely that high clouds in the atmosphere are preventing the observations from probing deep enough into the atmosphere to detect the water. Another explanation is that the water simply isn't there (which the theorists would argue is not a viable explanation). Also, it could be that the planet's atmosphere is isothermal (same temperature at all altitudes) which would preclude the formation of spectral features. In any case, the exciting nature of these observations points to the cutting edge state of this field. This technique could someday be used on transiting planets that are similar to the Earth. In this way, we might someday be able to detect the spectral signatures of life on other planets.
L.J. Richardson et. al. Nature 445, 892-895 (22 February 2007)