Exoplanet transit spectra with the Hubble Space Telescope


Two years ago, in a piece called “Transit spectra with the Hubble Space Telescope – the Dream”, we mused about the potential of our upcoming observations of hot Jupiters in transit spectroscopy with the Hubble Space Telescope. We had just obtained 124 orbits with HST to collect the transmission spectra of the atmospheres of ten hot Jupiters.

The observations are now completed, and the analysis of the results is advanced enough to compare dream and reality.

The results have not been disappointing. As happens with dreams, in many ways the actual spectra turned out to be very different from what we were expecting.

Here are the main results in a nutshell:

  1. there are lots of aerosols in most hot-Jupiter atmospheres. Be it clouds, haze, dust or droplets, these aerosols often leave clear marks on the transit spectra. Hazes like that first observed on HD 189733b are common, even on much hotter planets like WASP-12b. Given the temperatures (1000 to 3000 K), the hazes could be composed of silicates and iron at the cool end, and aluminium oxide at the hot end. Aerosols obscure the view to deeper layers and make the detection of molecules more difficult. This might be relevant in the quest to characterise Earth-like exoplanets. If high-altitude aerosols are also common in cooler atmospheres, they will add a challenge when attempting to measure the molecular signatures of life in exoplanet spectra.
  1. hot Jupiters are diverse. Every planet seems to look different in spectroscopy, even at similar temperature. We expected hot Jupiter spectra to line up in a neat temperature sequence, the way stars do, but the observations tell another story. In particular, there is no sign of the titanium oxide absorption expected in hot Jupiters above 1500 K. The effect of titanium oxide on spectra is so large that it  would have been easily detected. The fate of titanium oxide on hot Jupiters is a bit mysterious; it might condense on the cooler night side and sink as solid grains to the interior of the planet, clearing the atmosphere of titanium entirely.

The sketch below shows ten exoplanets, with temperature on the horizontal axis. The vertical axis characterises the spectrum as dominated by titanium oxide, sodium, clouds, or haze. The shaded area shows where we expected hot Jupiters to line up in this plot prior to the observations.

Do you see any pattern in this plot? If you do, contact me!

Simplified results from Hubble spectra of hot Jupiters, showing that (1) many hot Jupiters are shrouded in aerosols (2) titanium oxide is not observed (3) hot Jupiters are a diverse family.

Simplified results from Hubble spectra of hot Jupiters, showing that (1) many hot Jupiters are shrouded in aerosols (2) titanium oxide is not observed (3) hot Jupiters are a diverse family.


Four suns seen from within the atmosphere of their planet, to scale. Hot Jupiter sunsets are either blue-green (clear with sodium absorption), red-orange (due to scattering by fine glass grains), gray-white (scattering by larger or darker grain grains). [See these posts on the sunsets of WASP-12, HD 189733 and HD 209458]

Further details

Transit spectra covering wavelengths from 350 nm to 1 micron (near UV to near IR) were measured with the STIS instrument on the Hubble Space Telescope, in a programme led by David Sing at the University of Exeter.

A similar number of hot Jupiters was also observed in the 1-2 micron range with the WFC3 instrument on the HST, in a programme led by Drake Deming at the University of Maryland.


About Author

I am a professor of planetary science at the University of Exeter. My specialty is the study of exoplanets, in particular the observation and modelling of exoplanet atmospheres. I have done my PhD a the University of Geneva and worked in Chile, France and Switzerland.