Monday, December 14, 2015

Clear to Cloudy

As of this morning the NASA Exoplanet archive lists the discovery of 1916 exoplanets. These planets range in size and mass all the way from super Jupiters (over 10 times the mass of Jupiter), to Neptune’s, super-Earths (up to 10 times the mass of the Earth), and potentially Mars-sized or even Mercury-sized worlds. One of the most interesting discoveries in the search for exoplanets is that many of them are nothing like what we have in our solar system.

One such class of planet that was discovered is the hot Jupiter. Hot Jupiters are similar to Jupiter in mass , but can range in size from 0.8 - 2 times as large in their radius. Hot Jupiters also orbit much closer to their host stars than Jupiter in our own solar system, which sits at a cool 5 time the distance from the sun as the Earth, in fact hot Jupiters orbit 20 times closer to their star than we do to the Sun, which is even closer than even Mercury. This means that they are also tidally locked, with one face of their planet in constant daylight and the other in constant night.  

But now the new and exciting stuff...

To get an idea about what these strange new alien worlds are like, an international team of astronomers harnessed the observing power of the Hubble space telescope, and the Spitzer space telescope to conduct the most extensive study thus far to characterise the atmospheres of ten hot Jupiters. By looking at the stars light as the planet transits in front of the star, as Venus does in our solar system, from the perspective of the Earth we can detect the unique fingerprints of different molecules in the planets upper atmosphere as the light is filtered through before reaching our telescopes. This measurement of how the atmosphere absorbed light at different wavelengths is called transmission spectroscopy.

HST/Spitzer transmission spectral sequence
of hot Jupiter survey targets.
The solid lines are the model spectra fit to the measured transmission
spectra of each hot Jupiter in the survey, which are shown as data
points with their measured uncertainties. The spectra are offset and ordered
from top to bottom with low to high values of 
In the past only a small number of well-studied planets (HD 209458b, HD189733b, GJ 1214b) have been analyzed over a small wavelength range between 1.1-1.7 microns, just beyond the red part our eyes can see (Deming et al. 2013; Line et al. 2013; Sing et al. 2014; McCullough et al. 2014; Sing et al. 2015). This is an important wavelength range as it is where water vapor absorbs sunlight. Recent studies have shown that hot Jupiters have much smaller and muted water features than is predicted. The smaller signals could potentially be explained by lower amounts of water than was expected (Seager et al. 2005; Madhusudhan et al. 2014), which is a sign that water is removed somehow from the protoplanetary disk from which a solar system would form (Oberg, Murray-Clay, & Bergin 2011), but it is not clear if this high level of depletion is even possible. Alternatively, the weak signals of water vapor could be the result of clouds or hazes in the hot Jupiters atmosphere which obscure the water signals (Pont et al. 2013; Sing et al. 2014, 2015; Nikolov et al. 2015)

This new study published in Nature shows the measurements of ten hot Jupiter atmospheres from the optical (0.3 microns) all the way out into the mid-infrared (5 microns). This allows us to spectral resolve not only the water feature in the near-IR, but also the optical scattering and the IR molecular absorption features. The result of the study reveal a diverse group of planetary environments from clear atmospheres, which exhibit little to no evidence for high altitude clouds or hazes, revealing broad atomic signatures in the optical to large amplitude water features in the IR, to cloudy and hazy planets, which have strong Rayleigh scattering slopes in the optical and muted or even absent absorption features from water in the IR.

Pressure-Temperature profiles and condensation curves.P-T profiles were calculated from 1D radiative transfer
models (Fortney et al. 2008). Condensation curves are calculated for
chemical species expected in hot Jupiter atmospheres (Morley et al. 2012).
The presence of a cloud or haze can be predicted by the atmospheres temperature profile with altitude. Models show that as we move higher and higher in the planets atmosphere, decreasing log(pressure), the temperature decreases. Where the temperature and pressure of the atmosphere match the temperature and pressure where a material will change from a gas to a liquid is called the condensation point (Wakeford et al 2015). This will form the base of the cloud in that atmosphere. This is shown in the figure where the colored lines cross the grey dashed lines. Yet strangely the presence of obscuring clouds and hazes in the observations do not match up perfectly with the predictions from the models. So temperature and pressure alone cannot explain the clear to clouds atmospheres observed, suggesting circulation, vertical mixing need to be considered to transport heat and particles around the planet (Showman & Polvani 2011).

In the study we present a metric to distinguish between planetary atmosphere, comparing the difference between the measured planetary radius in the near-IR to that measured in the mid-IR (ΔZJ-LM), and correlate this with the strength (amplitude) of the water absorption feature. We additionally define the difference between the optical and mid-IR (ΔZUM-LM) that compares the strength of scattering in the optical, to the molecular absorption measured by Spitzer.

Transmission spectral index diagram: ΔZJ-LM vs H2O amplitude.The uncertainties represent the 1-sigma uncertainty. The purple and grey lines show model 
trends for cloudy and hazy atmospheres. We also show the clear-atom models with sub-solar
abundances in red. Note: WASP-39b and WASP-6b are missing as they currently have no
HST/WFC3 data. However, we have both scheduled for this coming year.
By plotting our measurements from this broad wavelength study we show that the results favor the obscuration of water by clouds or hazes rather than models where the water abundance is lower. The effect of clouds or hazes in the atmosphere reduces the water absorption amplitude, while raising the level of the near-IR to mid-IR continuum levels, leading to high ΔZJ-LM with low water amplitudes.

Theoretical model transmission spectra.
Model spectra assuming 1200K temperature with gravity of 25 m/s^2.
Models in each panel are compared to a clear, solar metallicity atmo (black).
This has revealed a continuum of hot Jupiter atmospheres from clear to cloudy. With the cloud formation in the atmospheres of these planets being highly sensitive to the temperature and pressure profile such steep gradients in the upper atmospheres, mean changes of just 100 degrees either way can totally change the cloud properties being observed. You can also see the different models that go with these metrics below.

This is the first time a large comparative study has been conducted on exoplantary atmospheres using consistent analysis techniques to combine datasets from different instruments and spacecraft. It is fantastic to finally see them all together in one place so that we can start to unravel the mysteries of these planets from afar.

The exciting part is now with this metric we can attempt to classify such strange alien planets, and the James Webb Space Telescope, Hubble’s successor, will expand our range of planets to study even further. Importantly, with more and more observations we may be able to make our own solar system planets part of this interstellar comparison. 

You can find the full nature paper on the Nature site here 

Also check out all of the press releases on the survey below