How to determine the atmospheric composition of an extrasolar planet

[Quantum]

Using a telescope at the European Southern Observatory in Chile, the researchers were able to study the planet by watching how it blocked some of the light of its host star as it passed in front of it.
"It makes the star look a little bit fainter - and it's actually a very good way of finding transiting planets - it's how this one was found," said Dr Southworth.
But different molecules in a planet's atmosphere - if it has one - absorb light in different ways, allowing scientists to look for their chemical signatures when the world transits its star.
The observations of planet GJ 1132b suggest that it has a thick atmosphere containing either steam and/or methane.
(Atmosphere found around Earth-like planet GJ 1132b
http://www.bbc.co.uk/news/science-env...)

[Matthew]

Hey, I just did a story on this! :) And yes, if the planet has a particularly dense atmosphere, it will absorb more in the way of IR light and appear larger in that wavelength than in others. Astronomers are also able to use spectroscopy to analyze the composition of an atmosphere, but its mainly been done so far with particularly large worlds that have had wider orbits from their star since they are easier to spot.

Upcoming missions like the James Webb Space Telescopes, and the New Worlds Mission will make it easier. Then we ought to be able to confirm the existence of atmospheres around smaller planets with smaller orbital periods more often.

https://exoplanets.nasa.gov/resources...
https://jwst.nasa.gov/

[Ian]

Matthew has one method, but it is not very reliable because the whole thing is at the edge of observational ability now. The James Webb telescope offers a better chance because it (I understand) can record infrared light and the spectral distribution. Specific molecules absorb particular infrared bands, thus methane exhibits C-H stretch and bending modes. CO2 does the same, except in different spectral positions, cyanides and acetylene completely different from most others, so if you can record a spectrum, you have the data. So far we cannot, and additionally this is far better done in space because our atmosphere contains IR absorbers. If the planet passes in front of the star, you can in principle detect what has been absorbed. An alternative is if the planet is to one side, and the star is blocked (in which was you get an emission spectrum - it is the same in a sense, but "upside down", i,e, you only detect the emission stretches.

It will always be difficult because dust tends to spoil everything. Also some molecules of interest may not absorb. To get a transition there has to be a change in dipole moment with the stretch. Thus a molecule like O2 or N2 or H2 won't give a signal (although they do with dust if adsorbed on it).