An outline of the paper: THE DEEP BLUE COLOR OF HD 189733b. by Evans et al.
Contributors to this website have recently participated in reaching another interesting milestone in the study of exoplanet atmospheres. Pointing the Hubble Space Telescope on HD 189733b, we have measured the reflection albedo of the atmosphere of this hot Jupiter in the visible; in other words, we have measured its actual colour, as perceived by the human eye.
The best fit to the data indicates that ‘189 is dark blue, as blue as Earth’s oceans seen from space. Of course, this result comes with the usual warnings: measurements of exoplanet atmospheres are always difficult because they must be teased out of the much larger signal from the parent star.
We find that the atmosphere of the planet reflects around 40% of the blue light from its parent star, and less than 10% of the red light, which makes it look very blue.
Given what we think we know about this planet and hot Jupiter atmospheres in general, we think that this is probably due to clouds of silicate grains scattering the light in the blue, and sodium absorbing the orange and red light. This is merely the most straightforward scenario, and with only two values measured in the reflection spectrum, we cannot be sure.
The albedo spectrum of the planet in the visible was measured by detecting the occultation of the planet by the star (the “secondary eclipse” at the opposite phase of the planetary transit). The dimming during the occultation is proportional to the brightness of the planet at full phase. We observed a dimming of 0.13 ± 0.03% in the blue and 0.00 ± 0.01% in the green and yellow.
This is the same method as used to measure the emission spectrum of transiting planets in the infrared. The occultation is much deeper in the infrared, but this is partly compensated by the fact that the star is brighter altogether in the visible, and that CCD cameras are more sensitive than infrared arrays.
Note that the deep blue colour of ‘189 is coherent with the “red sunset of ‘189” result from the transit spectrum. If red light from the star is absorbed by sodium and dust scatters blue light, the atmosphere will redden the light shining through it, but will appear blue in reflected light – just like the opalescent glass on the Wikipedia page for Rayleigh scattering, reproduced below.
The planet ‘189 is a specially favourable case, so it might be difficult to repeat this feat with other transiting planets or to get much more precise measurements in the near future. One catch is that the coming space telescope successor (JWST), as well as space missions dedicated to the measurement of exoplanet atmospheres, will be equipped with infrared cameras only, because they concentrate on the detection of molecules rather than clouds.
The main scientific implication of this result is that clouds seem to be important in the atmosphere of ‘189, as suggested already by the transmission spectrum. But the clouds are not high enough to dominate sodium absorption entirely and to send a large fraction of the incoming star light back so space.